guide on poverty measurement · 2018-02-14 · chapter 1 is an introduction describing the need for...
TRANSCRIPT
G
uide
on
Pove
rty
Mea
sure
men
t
Guide on Poverty Measurement
Palais des NationsCH - 1211 Geneva 10, SwitzerlandTelephone: +41(0)22 917 44 44E-mail: [email protected]: http://www.unece.org
Information ServiceUnited Nations Economic Commission for Europe
UN
ECEU
NITED
NATIO
NS
In the UNECE region, countries’ approaches to poverty measurement vary signi�cantly. There is a large spectrum of poverty indicators, wide varieties of de�nitions, methods, thresholds and data sources that are not fully matched by national or international guidelines.
This publication provides guidance on applying various measurement approaches and aims to improve the international comparability of poverty statistics.
Chapter 1 is an introduction describing the need for guidelines on poverty measurement and how is poverty measured today.
Chapter 2 provides an overview of poverty and related concepts such as inequality, social inclusion, vulnerability to poverty, and poverty risk.
Chapter 3 addresses the monetary approach to poverty, including the income and consumption expenditure measures that are most commonly used in measuring monetary poverty.
Chapter 4 introduces non-monetary deprivations, re�ecting Agenda 2030’s recognition that poverty is a multidimensional phenomenon.
Chapter 5 addresses the measurement of multidimensional poverty and demonstrates its relevance for policy design and analysis at global, regional and national levels.
Chapter 6 entitled “Challenges for the future” presents an overview of the areas envisaged for further work.
The publication mainly targets national statistical authorities and provides useful information for policymakers, researchers and other users of poverty data.
G
uide on Poverty Measurem
ent
G
UNIT
Guide
TEDNATIO
eon
ONSECON
Pove
UniNewYork
NOMICCO
erty
tedNatkandGene
MMISSION
Mea
ions eva,2017
NFOREUR
sure
ROPE
emennt
Note
The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of the Secretariat of the United Nations concerning the legal status of any country, territory, city or area, or of its authorities, or concerning the delimitation of its frontiers or boundaries.
Copyright © United Nations, 2017 All rights reserved worldwide
United Nations publication issued by the Economic Commission for Europe
ECE/CES/STAT/2017/4
iii
Preface
Poverty is increasingly recognized as a global phenomenon. The call for internationally comparable poverty measures is especially strong in the context of the 2030 Agenda for Sustainable Development. Moreover, the recent economic crisis has heightened the need for reliable and timely statistics for international monitoring and national policymaking on poverty reduction.
To improve the international comparability and availability of statistics on poverty and the related metadata, the Conference of European Statisticians established a Task Force in 2014. The Task Force on Poverty Measurement worked through 2015 and 2016 to develop the present Guide.
The Guide refers to the Sustainable Development Goals indicators and their underlying data needs and includes specific recommendations to national statistical offices. The Guide is based on the experience of UNECE member countries and other countries participating in the work of the Conference of European Statisticians.
The implementation of the Guide’s recommendations would improve international comparability of poverty statistics. The publication mainly targets national statistical authorities and provides useful information for policymakers, researchers and other users of poverty data.
UNECE is grateful to all experts who were involved in the preparation of this Guide.
iv
Acknowledgements
This Guide has been prepared by the UNECE Task Force on Poverty Measurement, which consisted of the following members: Giorgi Kalakashvili (National Statistics Office of Georgia), Bernd Becker (Federal Statistical Office, Germany), Nicoletta Pannuzi (Istat, Italy), Anna Szukiełoć‐Bieńkuńska (Central Statistical Office of Poland), Natalia Ignatova and Tatiana Velikanova (Rosstat, Russian Federation), Zeynep Gürsoy and Yakut Yüksel (Turkish Statistical Institute), Richard Tonkin (United Kingdom Office for National Statistics, ONS), John Shale and Jeetendra Seera (United Kingdom Department for Work and Pensions), Valentina Bryseva (CIS‐STAT), Didier Dupré, Jean‐Louis Mercy, Emilio Di Meglio and Jakub Hrkal (Eurostat), Marco Mira d’Ercole and Carlotta Balestra (OECD), Ben Slay, Elena Danilova‐Cross and Mihail Peleah (UNDP Istanbul Regional Hub), Joanne Bosworth (UNICEF), João Pedro Lopez Azevedo and Minh Cong Nguyen (World Bank), Sabina Alkire, Adriana Conconi and Bilal Malaeb (Oxford Poverty and Human Development Initiative, (OPHI)), Andres Vikat and Vania Etropolska (UNECE). The Task Force acknowledges input from Matthias Till (Statistics Austria).
The chapters of the Guide have been discussed and agreed by the entire Task Force. Some organizations took primary responsibility for drafting certain chapters, as follows: chapters 1 and 2 by UNDP and UNECE, chapter 3 by the United Kingdom (ONS), and chapters 4 and 5 by OPHI. Ben Slay (UNDP) edited the entire Guide.
UNECE acknowledges financial support from the United Nations Development Account (9th tranche) under the project “Promoting equality: strengthening the capacity of select developing countries to design and implement equality‐oriented public policies and programmes” (1415BG).
Cont
List of T
List of F
List of B
1 Int
1.1
1.2
1.3
1.4
1.5
1.6
1.7
2 Con
2.1
2.2
2.3
2.4
3 Mo
3.1
3.2
tents
Tables........
Figures ......
Boxes ........
roduction ..
1 Why this G
2 Objective
3 Outline of
4 Why mea
5 Poverty a
6 Monitorin
7 Internatio
nceptual Ba
1 The conce
2 The evolu
Ph2.2.1 Ba2.2.2 Re2.2.3 We2.2.4 Tim2.2.5 Mu2.2.6
3 Methodol
4 Measurem
No2.4.1 Dis2.4.2 Eq2.4.3 Po2.4.4 Re2.4.5 Me2.4.6
onetary Pov
1 Concepts
Int3.1.1 Un3.1.2 Un3.1.3 Ho3.1.4 Po3.1.5 Dis3.1.6
2 Welfare m
Inc3.2.1
..................
..................
..................
..................
Guide? .......
of the Guid
f the Guide
sure povert
nd the Mille
ng the Susta
onal compar
ackground .
epts of pove
tion of pov
hysical subsiasic needs ...elative deprealth ..........me poverty ultidimensi
logical choic
ment issues
on‐coveragesaggregatiouivalence soverty dynameporting on easuring no
verty ..........
and metho
troduction .nit of observnit of analysousehold deopulation cosaggregatio
measures ....
come conce
..................
..................
..................
..................
...................
de ...............
..................
ty? How is i
ennium Dev
ainable Dev
rability—ke
..................
erty, inequa
erty measu
istence ..........................ivation ............................................onal povert
ces .............
..................
e .................on ...............scales .........mics ...........poverty anon‐monetar
..................
ods ..............
...................vation ........sis ...............efinition .....overage ......on ...............
...................
epts and de
v
..................
..................
..................
..................
...................
...................
...................
t measured
velopment
elopment G
ey to succes
..................
ality, and so
rement ......
...................
...................
...................
...................
...................ty................
...................
...................
...................
...................
...................
...................d inequalityry poverty ..
..................
...................
...................
...................
...................
...................
...................
...................
...................
finitions .....
..................
..................
..................
..................
...................
...................
...................
d today? .....
Goals .........
Goals ..........
sful policies
..................
ocial exclusi
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................y ....................................
..................
...................
...................
...................
...................
...................
...................
...................
...................
...................
..................
..................
..................
..................
...................
...................
...................
...................
...................
...................
s .................
..................
on ..............
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
..................
...................
...................
...................
...................
...................
...................
...................
...................
...................
.................
.................
.................
.................
..................
..................
..................
..................
..................
..................
..................
.................
..................
..................
..................
..................
..................
..................
..................
..................
..................
..................
..................
..................
..................
..................
..................
..................
.................
..................
..................
..................
..................
..................
..................
..................
..................
..................
........ viii
.......... ix
........... x
........... 1
............ 1
............ 2
............ 2
............ 3
............ 5
............ 6
............ 8
......... 10
.......... 10
.......... 11
.......... 11
.......... 12
.......... 12
.......... 14
.......... 14
.......... 16
.......... 18
.......... 20
.......... 20
.......... 20
.......... 20
.......... 21
.......... 21
.......... 24
......... 25
.......... 25
.......... 25
.......... 26
.......... 27
.......... 27
.......... 28
.......... 32
.......... 34
.......... 34
3.3
3.4
3.5
3.6
4 Pov
4.1
4.2
4.3
4.4
Ad3.2.2 Da3.2.3 Co3.2.4 Co3.2.5 Da3.2.6 Us3.2.7 Ke3.2.8 We3.2.9
3 Setting a p
Ab3.3.1 Ab3.3.2 Re3.3.3 Ke3.3.4
4 Poverty in
Ov3.4.1 Sta3.4.2 Dy3.4.3
5 Improving
Po3.5.1 Mo3.5.2 Me3.5.3
6 Summary
Un3.6.1 Dis3.6.2 We3.6.3 Po3.6.4 Ind3.6.5 Re3.6.6 Me3.6.7
verty Dashb
1 Introducti
2 Processes
3 Comparab
Re4.3.1 Ex4.3.2 Ex4.3.3 As4.3.4
4 Material d
Re4.4.1 Ex4.4.2 Alt4.4.3 Ca4.4.4 As4.4.5
dvantages aata sources onsumptiononsumptionata sources sing multipleey measuremealth ..........
poverty line
bsolute and bsolute poveelative Poveey issues .....
ndicators ....
verview ......atic measurynamic mea
g internatio
overty and tonetary povetadata con
of recomm
nit of observsaggregatioelfare measoverty line ..dicators .....egional poveetadata ......
boards and
ion .............
s and princip
ble dashboa
equirementsamples of cample of dasessments
deprivation
equirementsample: Matternative mse study: Ensessment o
nd disadvanfor househ expenditur expenditurfor consume welfare mment issues...................
e .................
relative poerty lines ...erty Lines .......................
...................
...................res ..............asures .........
nal compar
he SDGs ....verty indicansiderations
mendations .
vation/analon of data ...sures ................................................erty measur...................
the Materi
...................
ples ............
ards ............
s for compacomparableashboards iof compara
indices ......
s .................terial depriv
material depnglish indexof material d
vi
ntages of inold income re: conceptre: advantamption expemeasures ....s ....................................
...................
verty lines ..........................................................
...................
...................
...................
...................
rability: Reg
...................ators for CESs .................
...................
ysis and po............................................................................res ................................
ial Deprivat
...................
...................
...................
arable dashe internationn Europe ...able dashbo
...................
...................vation in Eurivation ratx of multipledeprivation
ncome as a ..................s and definges and disnditure ...............................................................
...................
...................
...................
...................
...................
...................
...................
...................
...................
gional indica
...................S countries ...................
...................
opulation co..................................................................................................................
tion Indices
...................
...................
...................
boards .......nal dashboa...................oards...........
...................
...................urope ..........te ................e deprivatio indices ......
poverty me...................itions .........advantages............................................................................
...................
...................
...................
...................
...................
...................
...................
...................
...................
ators and m
........................................................
...................
overage .........................................................................................................................
s ................
...................
...................
...................
...................ards: MDGs......................................
...................
...................
...................
...................on ..................................
easure ............................................s.........................................................................................
..................
..................
..................
..................
..................
..................
..................
..................
..................
metadata ....
..................
..................
..................
..................
..................
..................
..................
..................
..................
..................
..................
.................
..................
..................
..................
.................. and SDGs .....................................
..................
..................
..................
..................
..................
..................
.......... 38
.......... 40
.......... 43
.......... 44
.......... 45
.......... 46
.......... 49
.......... 61
.......... 63
.......... 63
.......... 63
.......... 72
.......... 78
.......... 85
.......... 85
.......... 86
.......... 91
.......... 95
.......... 95
.......... 96
.......... 97
.......... 99
.......... 99
.......... 99
........ 100
........ 101
........ 102
........ 102
........ 102
....... 104
........ 104
........ 104
........ 106
........ 106
........ 107
........ 107
........ 109
........ 111
........ 112
........ 112
........ 115
........ 118
........ 121
4.5
5 Mu
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8
5.9
6 Cha
6.1
6.2
6.3
6.4
6.5
6.6
6.7
6.8pov
6.9
Referen
Annex Ifor Sust
Annex I
5 Conclusio
ultidimensio
1 Introducti
2 Overview
3 Requirem
4 Steps to b
Pre5.4.1 Ide5.4.2 Ag5.4.3
5 Key decisi
Un5.5.1 Dim5.5.2 Ind5.5.3 We5.5.4 Po5.5.5 Sh5.5.6
6 Case stud
7 Case stud
8 Case stud
9 Assessme
Ad5.9.1 Dis5.9.2
allenges for
1 Hard to re
2 Imputed r
3 Social tran
4 Wealth....
5 Comparab
6 Comparab
7 Individual
8 Spatial difverty ..........
9 Subjective
nces...........
I Goal 1 tainable De
II Results
n ................
onal Povert
ion .............
..................
ments ..........
build an MP
eliminary stentification ggregation ..
ions ............
nit of identifmensions ...dicators andeights ........overty cut‐oould incom
y: The OPH
y: Colombia
y: The Mex
nt of the M
dvantages osadvantage
r the Future
each popula
rent and ho
nsfers in kin
..................
ble welfare
bility of mu
l level pove
fferences w..................
e poverty ...
..................
and Goal 1evelopment
s of the UNE
...................
ty Indices ...
...................
...................
...................
I .................
tep ............. of the poo...................
...................
fication .........................d deprivatio...................off(s) ...........me be includ
I/UNDP Glo
a’s national
ican MPI ....
MPI ..............
of the MPI ...es ................
e ................
ations .........
ousing cost .
nd ...............
...................
aggregates
ltidimensio
rty measure
within count...................
...................
..................
0 poverty‐rt ................
ECE survey
vii
...................
..................
...................
...................
...................
...................
...................r ....................................
...................
...................
...................on cut‐offs .......................................ed in the M
obal MPI .....
MPI–Struc
...................
...................
...................
...................
..................
...................
...................
...................
...................
s ..................
nal poverty
es ...............
ries with re...................
...................
..................
related targ..................
on poverty
...................
..................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................MPI? ............
...................
ture and po
...................
...................
...................
...................
..................
...................
...................
...................
...................
...................
y measures .
...................
egard to con...................
...................
..................
gets and ind..................
y measurem
...................
..................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
...................
olicy applica
...................
...................
...................
...................
..................
...................
...................
...................
...................
...................
...................
...................
nsumption a...................
...................
..................
dicators in t..................
ment ...........
..................
.................
..................
..................
..................
..................
..................
..................
..................
..................
..................
..................
..................
..................
..................
..................
..................
ations ........
..................
..................
..................
..................
.................
..................
..................
..................
..................
..................
..................
..................
and income..................
..................
.................
the 2030 Ag.................
.................
........ 122
....... 123
........ 123
........ 124
........ 125
........ 126
........ 126
........ 127
........ 127
........ 130
........ 130
........ 130
........ 132
........ 133
........ 135
........ 136
........ 140
........ 143
........ 146
........ 147
........ 147
........ 154
....... 156
........ 156
........ 158
........ 158
........ 159
........ 160
........ 161
........ 161
e ........ 162
........ 162
....... 164
genda ....... 178
....... 181
viii
ListofTables
Table 2.1 Comparing the definitions of social exclusion and poverty ............................... 11 Table 2.2 Reporting on MDG indicators in international and national databases ............ 22 Table 2.3 Shares of population living below income poverty lines, for selected
countries, 2009‐2012 .......................................................................................... 23 Table 3.1 Income components in the conceptual definition of the Canberra Group
Handbook............................................................................................................ 36 Table 3.2 International poverty lines by group .................................................................. 72 Table 3.3 Relative poverty risk of children vs the elderly, by type of equivalence
scale .................................................................................................................... 79 Table 5.1 The dimensions, indicators, deprivation cut‐offs and weights of the
Global MPI ........................................................................................................ 140 Table 5.2 Average deprivation in pair‐wise indicators across 101 developing
countries ........................................................................................................... 150
ix
ListofFigures
Figure 2.1 Numbers of persons at risk of poverty or social exclusion by type of risks, EU‐28, 2015 ........................................................................................................ 14
Figure 3.1 Net wealth of private households in Austria, 2014 ............................................ 61 Figure 3.2 At‐risk‐of‐poverty rates anchored at a fixed moment in time (2008)
versus “standard” at‐risk‐of‐poverty rates, 2008‐2014 ..................................... 77 Figure 4.1 Severe material deprivation rate, by sex and age group, EU 28, 2010 and
2013 .................................................................................................................. 114 Figure 4.2 Severe material deprivation rate by household type, educational
attainment, and country of birth, EU‐28, 2013 ................................................ 114 Figure 4.3 Map of severely materially deprived countries, 2014, per cent of
population ......................................................................................................... 115 Figure 4.4 Overview of the English multiple deprivation methodology ........................... 119 Figure 4.5 The Index of Multiple Deprivation (IMD), 2015 ............................................... 120 Figure 5.1 Steps for measuring a Multidimensional Poverty Index .................................. 128 Figure 5.2 Moving from individual deprivations to the deprivation score ....................... 141 Figure 5.3 Dimensions, Variables and Weights for the MPI Colombia ............................. 144 Figure 5.4 Poverty identification by the National Council for the Evaluation of Social
Development Policy (CONEVAL) ....................................................................... 147
x
ListofBoxes
Box 2.1 Time poverty—country examples ......................................................................... 15 Box 3.1 Collecting data on the homeless in Italy ............................................................... 29 Box 3.2 UNDP’s experience collecting Roma poverty data ................................................ 31 Box 3.3 Persons at risk of poverty and beneficiaries of social transfers: Different
concepts ‐ different people? A case study from Germany 2014 .......................... 37 Box 3.4 The combined use of survey and administrative data .......................................... 40 Box 3.5 Comparing poverty estimates using income, expenditure, and material
deprivation data .................................................................................................... 47 Box 3.6 Imputed rent in EU‐SILC: 2007‐2010 ..................................................................... 52 Box 3.7 Remittances and poverty in Eastern Europe and Central Asia.............................. 55 Box 3.8 Poverty measures including social transfers in kind in United Kingdom and
Finland ................................................................................................................... 57 Box 3.9 The measurement of imputed rents and social transfers in kind in the OECD ..... 59 Box 3.10 The official thresholds measure in the United States ........................................... 68 Box 3.11 Establishing an absolute poverty line in Italy ........................................................ 69 Box 3.12 Establishing an absolute poverty line in the Russian Federation .......................... 70 Box 3.13 Use of mean and median income in the at‐risk‐of‐poverty threshold in EU‐
SILC ........................................................................................................................ 74 Box 3.14 EU‐SILC use of different relative poverty thresholds ............................................ 75 Box 3.15 Influence of different equivalence scales on poverty rates in Poland .................. 81 Box 3.16 Equivalisation in Russian poverty measurement .................................................. 82 Box 3.17 Poverty indicators in Russian Federation .............................................................. 88 Box 3.18 Persistent poverty in the United Kingdom and EU ................................................ 92 Box 3.19 Poverty entry and exit rates in EU countries ......................................................... 94 Box 4.1 EU‐SILC and the Open Method of Coordination ................................................. 105 Box 4.2 European social indicators ................................................................................... 107 Box 4.3 Measuring basic deprivations in the Russian Federation.................................... 109 Box 4.4 Material deprivation in the Republic of Moldova ............................................... 117 Box5.1 AnexampleofhowtocomputetheMPI .......................................................... 129 Box 5.2 The choice of dimensions in different national MPIs .......................................... 131 Box 5.3 Multidimensional Poverty Indicators in Europe: EU‐SILC ................................... 134 Box 5.4 Towards a multidimensional poverty index in Germany .................................... 138 Box 5.5 An MPI for Kyrgyzstan ......................................................................................... 142 Box 5.6 Lessons from Colombia’s National MPIs ............................................................. 145 Box 5.7 Subgroup decomposition and dimensional contribution—Pakistan .................. 151 Box 6.1 Household surveys in Ukraine ............................................................................. 157
1
1 Introduction
1.1 WhythisGuide?
1. Poverty is increasingly recognized as a global phenomenon. The call for internationally comparable poverty measures is especially strong in the context of the 2030 Agenda for Sustainable Development. Moreover, the recent economic crisis has heightened the need for reliable and timely statistics for international monitoring and national policymaking on poverty reduction.
2. In the UNECE region, countries’ approaches to poverty measurement vary significantly. For many indicators, wide varieties of definitions, methods, and primary data sources are not fully matched by national or international guidelines for their application.
3. The Rio Group created by the United Nations Statistical Commission published a Compendium of best practices in poverty measurement (Rio Group, 2006). While the Compendium presents important concepts and definitions, it also pointed out that the state of the art and the very unequal availability of statistical instruments across countries were not conducive to the preparation of a universally applicable handbook at that time.
4. The Canberra Group handbook on household income statistics (UNECE, 2011) presents the concepts and components of household income, describes country practices and provides guidance on quality assurance and dissemination. It also includes a brief section on the analysis of income poverty. The Canberra Group handbook thus addresses the methodological basis for income poverty measures, but does not specifically elaborate and provide recommendations on poverty indicators and the related methodological choices. Furthermore, issues of non‐monetary poverty were beyond the scope of the Handbook.
5. In 2013, the OECD published the Framework for Statistics on Distribution of Household Income, Consumption and Wealth (OECD, 2013a), which is an internationally agreed framework to support the joint analysis of micro‐level statistics on household income, consumption and wealth. The same year the OECD published also the Guidelines for Micro Statistics on Household Wealth, an internationally agreed set of guidelines for producing micro statistics on household wealth (OECD, 2013b).
6. In 2012, the Bureau of the Conference of European Statisticians (CES) conducted an in‐depth review of poverty statistics based on a paper by the State Statistics Service of Ukraine and Eurostat (UNECE, 2012a,b). The review provided an analysis of the methodological issues underlying poverty measurement and presented two case studies: one at international level (Eurostat) and the other at national level (Ukraine). As a follow‐up to the in‐depth review, the Bureau requested the CES secretariat to organize a seminar to discuss how to improve poverty measurement.
7. The seminar “The way forward in poverty measurement” was held in Geneva on 2‐4 December 2013 with representatives from 29 countries and major international agencies active in poverty measurement in the UNECE region (CISSTAT, Eurostat, OECD, UNDP, World Bank). Participants discussed the main methodological issues in poverty measurement, data comparability, and inter‐linkages between poverty, inequality, vulnerability and social exclusion. The seminar identified the need for guidelines and
Chapter1Introduction
2
recommendations for improving the international comparability and availability of poverty statistics, and recommended that a Task Force undertake this work.
8. The CES Bureau established the Task Force on Poverty Measurement in 2014. It worked through 2015‐2016 to develop this Guide.
1.2 ObjectiveoftheGuide
9. The objective of the Guide is to provide guidance in applying various measurement approaches at national level and to improve the international comparability of poverty statistics. The Guide focuses on areas where the statistical community has expressed a particular need for further guidance, which include availability and comparability of key poverty measures, data requirements and measurement issues, and recent approaches to poverty measurement.
10. The Guide refers to the Sustainable Development (SDG) indicators and their underlying data needs and includes specific recommendations to national statistical offices. The Guide is based on the experience of UNECE member countries and other countries participating in the work of the Conference of European Statisticians1.
11. The Guide is primarily aimed at statisticians. It may also be relevant for policymakers for formulating targets for poverty reduction.
1.3 OutlineoftheGuide
12. Chapter 2 provides an overview of poverty and related concepts such as inequality, social inclusion, vulnerability to poverty, and poverty risk. It discusses the importance of poverty measurement and opens the debate about the advantages and disadvantages of different approaches and the value of complementary measures. The chapter offers a synopsis of the methodological choices countries have and defines the bigger scope of measurement challenges in our contemporary world.
13. Chapter 3 addresses the monetary approach to poverty, including the income and consumption expenditure measures that are most commonly used in measuring monetary poverty. The chapter explains concepts and definitions, provides an overview of data sources, and discusses the advantages and disadvantages of various welfare measures. It examines in detail such key measurement issues as measuring self‐employment income, goods and services produced for own consumption, transfers between households, social
1 The Conference of European Statisticians is composed of national statistical organizations in the UNECE region (for UNECE member countries, see http://www.unece.org/oes/nutshell/member_states_representatives.html) and includes in addition Australia, Brazil, Chile, China, Colombia, Japan, Mexico, Mongolia, New Zealand, and Republic of Korea. The major international organizations active in statistics in the UNECE region also participate in the work, such as the statistical office of the European Commission (Eurostat), the Organization for Economic Cooperation and Development (OECD), the Interstate Statistical Committee of the Commonwealth of the Independent States (CIS‐STAT), the International Monetary Fund (IMF) and the World Bank.
Chapter1Introduction
3
transfers, and transfers in kind. The chapter also reviews various approaches to setting a poverty line or threshold, illustrated with country examples. It identifies policy‐relevant poverty indicators, concerning the level and depth of poverty, and how these change over time. Finally, the chapter provides an overview of current practices, highlighting challenges related to assuring the comparability of poverty estimates.
14. Today it is broadly recognised that poverty reaches beyond people’s material conditions, including aspects such as poor health, job insecurity, social exclusion, malnutrition, and lack of personal security. Moreover, an integrated measure of multidimensional poverty has been included in the SDGs, to complement income poverty measures and show interconnected deprivations. Chapter 4 therefore introduces non‐monetary deprivations, reflecting Agenda 2030’s recognition that poverty is a multidimensional phenomenon. The chapter starts by showing the motivations for multidimensional measurement, with an emphasis on European countries. It then shows how countries can design basic dashboards of social indicators and gives examples from the region. The chapter also introduces the indices of multiple deprivations and provides examples of material deprivation measures in Europe. On these topics the Guide does not provide specific recommendations. Nevertheless, the users may find useful the experiences that exist on some countries and organizations.
15. Chapter 5 addresses the measurement of multidimensional poverty and demonstrates its relevance for policy design and analysis at global, regional, and national levels. The key challenges faced by statistical offices in developing a multidimensional poverty index include identifying the various welfare dimensions, selecting indicators in assessing deprivations at the individual or household level, and fixing poverty lines both for each dimension and overall. Although these measures are in general adapted to national circumstances, the need to ensure comparability at global and regional levels is also recognized. The chapter describes relevant measurement experiences in other regions and provides guidance to countries interested in developing multidimensional poverty measures regarding measurement design and how measures guide policy.
16. On some topics, the Guide does not make any clear‐cut recommendations due to insufficient evidence from current practice. Such areas include the measurement and consideration in poverty estimates of social transfers in kind, household wealth, housing costs, and individual‐level poverty. Furthermore, a person’s own subjective perceptions of his or her well‐being are important for understanding poverty, and robust measures of this would need to be worked out. An overview of the areas envisaged for further work is provided in Chapter 6 entitled “Challenges for the future”.
17. Annex I presents poverty‐related targets and indicators in the 2030 Agenda for Sustainable Development. Results from the UNECE survey on methods of poverty measurement in official statistics are presented in Annex II.
1.4 Whymeasurepoverty?Howisitmeasuredtoday?
18. UNDP’s 1991 Human Development Report captured the human development paradigm in a single sentence: “The real objective of development is to increase people’s choices”. The underlying concept is the ability to live long, healthy, and creative lives.
Chapter1Introduction
4
Additional choices include political freedom, guaranteed human rights, and self‐respect—what Adam Smith called the ability to mix with others without being “ashamed to appear in public”. From this standpoint, poverty is the inability to obtain or realize choices and opportunities; it is a violation of human dignity. Poverty means a lack of basic capacity to participate effectively in society. It means not having enough to feed and clothe a family, not having a school or health clinic to go to, not having land on which to grow one’s food or a job to earn one’s living, not having access to credit. It means insecurity, powerlessness and exclusion of individuals, households, and communities. It means susceptibility to violence, and it often implies living in marginal or fragile environments, without access to clean water or sanitation (United Nations Economic and Social Council, 1998).
19. This broad definition of poverty should lend itself to practical measurement, which in turn should inform public discourse and policy actions. Poverty measurement therefore faces: (i) methodological issues (“get it right”); and (ii) public policy concerns (“make it useful”). This Guide addresses both sets of issues, to be useful for evidence‐based policymaking at global, regional, national, and even sub‐national levels. Such aspirations imply additional requirements for poverty indicators, as the meaning and measurement of poverty at these different levels can be quite different. While attempts to produce and use globally comparable poverty statistics inevitably face questions about different living standards and lifestyles, they make possible international comparisons and efforts to establish best practices. For these reasons, the promotion of international comparability is of great importance. Moreover, the adoption of Agenda 2030 further underscores the imperative of developing guidelines and identifying best practices in measuring international progress in poverty eradication.
20. Poverty should be measured for a number of different reasons. First, poverty measures provide estimates of the magnitude of the problem, and raise its visibility—they keep poor people on the policy agenda. Second, poverty measures are needed to identify poor people and pockets of poverty, and then to target appropriate policy interventions. This requires data disaggregation, in order to identify population groups that face higher risk of poverty, based inter alia on personal characteristics, family structure, place of residence, etc. It also requires dynamic measures that can monitor poverty over time and identify those trapped in poverty for longer periods. High quality poverty statistics are therefore needed to monitor and evaluate outcomes—especially the effectiveness of policy, programming, and project interventions focusing on poor people.
21. Poverty measurement has direct implications for policymaking, as different perspectives on poverty can produce different empirical conclusions. It starts with conceptual definition of what exactly is being measured. Are we concerned about inequality at the lower end of the distribution, falling short of some absolute minimum living standards, the inability to “keep up with the Joneses”, or some broader type of social exclusion? Once the basic conceptual questions are answered, the definition of poverty should be operationalized in statistical terms. This seemingly technical issue can have serious (but often hidden) implications for policies. For instance, the use of different equivalence scales can produce different results for child versus elderly poverty, which can in turn create mixed signals for social protection policies.
Chapter1Introduction
5
1.5 PovertyandtheMillenniumDevelopmentGoals
22. The Millennium Declaration was adopted by heads of State and Government at the United Nations General Assembly in 2000; the Millennium Development Goals (MDGs) were adopted soon after. The eradication of extreme poverty and hunger were at the top of the agenda, as reflected in MDG 1 “Eradicate extreme hunger and poverty”,2 which was supported by two targets: Target 1 (“Halve, between 1990 and 2015, the proportion of people whose income is less than $1 a day”, in purchasing‐power‐parity [PPP] terms) and Target 2 (“Halve, between 1990 and 2015, the proportion of people who suffer from hunger”). These two targets were supposed to be monitored at the global level by five indicators.3 However, recent studies show that numerous statistical offices were unable to collect, analyse, and disseminate the data used for MDG reporting. MDG statistics were often based on donor‐funded surveys or modelling exercises (Loewe and Rippin, 2016).
23. Criticisms to the “$1‐a‐day” poverty line stressed not only its arbitrariness, but also its failure to take into consideration other basic needs apart from food and essential non‐food spending, such as housing, clothing, and heating. In addition, the low $1‐per‐day poverty line was not relevant for many countries, which contributed to the slow take‐up of the MDG agenda in many countries.4
24. More broadly MDG progress was measured via over 60 internationally agreed (global) indicators; many others were used at the national level. With regards to poverty, the global MDG indicators were tailored to the specific situation of low‐income countries. In 1990, people classified as living in extreme poverty based on this threshold were mainly found in rural regions. Nowadays, one fourth of the extremely poor live in cities. Therefore, a number of countries set their own national targets and added alternative indicators in order to capture these trends.
25. The current globally used poverty threshold of PPP$1.90/day is also very low for countries in the UNECE region. To remedy this situation, the World Bank has suggested that in middle‐income countries, two or more thresholds should be used. Other issues with absolute poverty lines are apparent in their high sensitivity to the choice of the PPP base year, the exchange rate used to convert income in national currency into US dollars, and the basket of goods chosen to compute the PPP. Partly due to these problems, some institutions like the European Union and the OECD do not use absolute poverty thresholds for international comparisons, but rather rely on relative thresholds expressed as a share of median income (Bradshaw and Mayhew, 2011).
2 For the list of MDG goals see http://www.unmillenniumproject.org/goals/gti.htm#goal1. 3 The Target 1 indicators were: (1) the proportion of the population below living $1 (1993 PPP) per day (World Bank) (For monitoring country poverty trends, indicators based on national poverty lines should be used, where available.); and (2) the poverty gap ratio [incidence x depth of poverty] (World Bank). Target 2 indicators included: (3) the share of the poorest quintile in national consumption (World Bank); (4) the prevalence of underweight children under five years of age (UNICEF‐WHO); and (5) the proportion of the population below minimum dietary consumption levels (FAO). 4 Similar issues appeared for the global Multidimensional Poverty Index. The deprivation thresholds selected were quite demanding, resulting in very low multidimensional poverty headcounts for many countries in Europe and Central Asia.
Chapter1Introduction
6
26. A general pattern was that, while richer countries generally engaged in less detailed and comprehensive poverty reporting under MDG 1, they frequently added other “national” targets and indicators that better suited their circumstances. These included, for example, measures of poverty prevalence among ethnic minorities such as the Roma, single mothers, or the proportion of population that depends on social benefits.
1.6 MonitoringtheSustainableDevelopmentGoals
27. The Sustainable Development Goals (SDGs) were adopted in September 2015 by world leaders as the monitoring framework for the 2030 Agenda for Sustainable Development, a plan of action for “people, planet, peace, partnership, and prosperity”. Consisting of 17 goals and 169 targets, the SDGs build on the development journey inherited from the MDGs. Their reach is however much wider than poverty, gender, hunger, and major health problems. The SDGs break new ground by addressing inequalities, economic growth, decent jobs, energy, natural resources and environment, climate change, human settlements, and peace and justice, among others. They represent an agreed vision to put people and planet on a sustainable path by 2030.
28. There are a number of other important differences between the SDGs and the MDGs. First, while the MDGs were driven to a significant extent by the donor community,5 the SDGs were developed by all Member States through a participatory process. While the MDGs were applicable mostly to the least developed countries, the SDGs offer an agenda for all people of the world, putting specific emphasis on “leaving no one behind”—which has serious implications for monitoring and evaluation. SDG targets go beyond averages and refer to different groups (e.g., women and men; migrants; urban and rural inhabitants; the poor, middle‐class, and the more well off). Last but not the least, the SDGs offer an integrative agenda (compared with the narrower, sectorial MDGs).
29. These differences have important implications for SDG monitoring, and especially for poverty measurement. First, SDG goals and targets have to be treated as a network of targets, rather than as a list of standalone isolated variables (Le Branc, 2015).6 As a result, poverty‐related targets and indicators are found not only under Goal 1 (“End poverty in all its forms everywhere”) and Goals 10 (“Reduce inequality within and among countries”), but in many others goals (see annex I for overview of Goals 1 and 10 SDG targets and indicators). The set of “poverty” indicators relevant for these goals will therefore be much larger than for the SDGs, including both absolute poverty (Indicator 1.2.1: “The proportion of the population living below the national poverty line, by sex and age”), relative poverty (Indicator 10.2.1: “The proportion of people living below 50 per cent of median income, by age, sex and persons with disabilities”), non‐income poverty measures (Indicator 6.2.1: “The proportion of the population using safely managed sanitation services, including a hand‐washing facility with soap and water”), as well as multidimensional poverty (Indicator 1.2.2:
5 The goals of the “Shaping the 21st Century” report became the basis of the United Nations’ Millennium Declaration and its MDGs (Organisation for Economic Co‐operation and Development, 1996). 6 See also http://peleah.me/sdg/sdgs‐targets.html for SDGs as a Network of Targets.
Chapter1Introduction
7
“The proportion of men, women and children of all ages living in poverty in all its dimensions according to national definitions”).
30. This comprehensive set of indicators represents a challenge for monitoring. While a global list of indicators has been agreed (United Nations Statistical Commission, 2016a), many of these indicators either lack an established methodology (so‐called “tier 3 indicators”) or are not supported by the regular production of the relevant official statistical data (“tier 2 indicators”). Out of the 229 approved global SDG indicators, only 119 are at present classified as “ready to go” tier 1 indicators; 44 are classified as tier 2 indicators, and another 76 are tier 3 indicators. In response, the United Nations Statistical Commission has “emphasized that the global indicators . . . are intended for global follow‐up and review of the 2030 Agenda for Sustainable Development and are not necessarily applicable to all national contexts, and that indicators for regional, national and subnational levels of monitoring will be developed at the regional and national levels” (Ibid., paragraph 47/101/(i)).
31. The SDG goals and targets pose significant challenges for national statistical offices, in terms of the capacity to produce the data needed to use the required indicators. Some of the SDG indicators are currently set up in a very general form and countries will require further methodological guidance in order to produce the data needed for their use. An example of such broadly defined indicators are those for target 1.2 (“Reduce at least by half the proportion of men, women and children of all ages living in poverty in all its dimensions according to national definitions”); 1.2.1 (“The proportion of the population living below the national poverty line, by sex and age”); and 1.2.2 (“The proportion of men, women and children of all ages living in poverty in all its dimensions according to national definitions”).
32. Most SDG indicators required for monitoring poverty, inequality, and employment come from household surveys. However, many household surveys are not designed in order to measure living standards and poverty; the emphasis is instead on measuring food consumption, housing services, and the cost of living (Gibson, 2015). In addition, in many countries the surveys are conducted on an irregular basis. Even if countries have a regular household survey in place, the data provided by the survey may be either insufficient or not in line with the international standards. Sub‐national poverty measurement, in particular, faces issues of inadequate data, as in most cases surveys are not representative at the local level.
33. Although the development of the SDG indicators at the regional, national, and sub‐national levels will pose challenges for national statistical offices and international organizations, it could also offer opportunities to strengthen statistical systems and make better use of innovative and inclusive data techniques for monitoring sustainable development. UN‐led discussions for the Europe and Central Asia region concluded that building national ownership of SDGs in national policy frameworks would require tailoring indicators to national conditions (United Nations Development Programme Istanbul Regional Hub, 2016). To do this, governments would need to build new statistical partnerships for poverty reduction, improve metadata for survey quality, and assess data ecosystems. This Guide seeks to support these efforts.
Chapter1Introduction
8
1.7 Internationalcomparability‐keytosuccessfulpolicies
34. There is a large global spectrum of poverty indicators and definitions. To give a comprehensive picture of poverty, national statistical offices use multiple concepts and thresholds. In addition to absolute poverty lines, many countries use relative lines defined as a certain percentage of national median income. This is the most frequently used measure in richer countries in the region. However, in times of crisis, changes in the percentage of people living under such poverty line may lead to counterintuitive results, because the median income to which the line relates may fall by more than the incomes of the poorest households. Counterintuitive results could also be observed in times of economic growth when the benefits of growth are distributed unequally towards the rich, in which case the higher median would show increase in the number of poor. There is consensus that no single approach is sufficient for monitoring poverty at the national and regional levels. The results from different approaches thus have to be communicated clearly, to allow correct interpretation of the different measures.
35. For national governments, the availability of comparable poverty measures can provide important information when dealing with the implementation or evaluation of policies and programmes. Without international comparisons, it is difficult for countries to measure their progress towards eradicating poverty. In the UNECE region, this implies both comparing efforts to neighbours to establish best practices in the region, and developing key statistical measures to facilitate comparison across sub‐regions. It could also mean that lower middle‐income countries may need to compare their approaches and poverty conditions to those prevailing in more developed economies, in order to establish programmes based on what has been done in similar contexts.
36. Use of the same poverty definitions operationalized in different ways (e.g., by using different equivalence scales, or using income rather than consumption as a welfare metric) can produce quite different results, both within and across countries. This in turn can also affect national and regional policy decisions. Moreover, the choice of definitions and indicators for monitoring countries’ current state and progress faces certain trade‐offs. On the one hand, ensuring international comparability suggests the use of universal definitions and harmonised methodologies; but on the other hand, a certain degree of flexibility is needed for a measure to be truly meaningful in a country‐specific context—suggesting the use of indicators that reflect national characteristics. Countries should therefore measure poverty in ways that respond to their needs and policy priorities. To preserve this flexibility, keeping two poverty targets (global and national) for both monetary and multidimensional poverty is foreseen in SDG monitoring, as was the case with MDGs.
37. Many international organizations—the World Bank, OECD, UNDP, Eurostat, just to mention a few—produce poverty data. There have been continuous efforts to improve capacity in statistical offices to develop poverty measures in line with international standards. However, in most cases, these data are not comparable and often cover only a limited number of countries. A lack of comparable data across countries and time impedes effective policy actions. Data produced by countries are not always comparable internationally, largely for two main reasons:
Country data primarily respond to national needs, which do not always correspond to international standards; and
Chapter1Introduction
9
Country data reflect national statistical capacities, which are not always able to meet international standards.
38. Both of these concerns are relevant for the UNECE region—which is quite heterogeneous in terms of development levels, so that countries have different needs when measuring poverty. The periodicity of surveys providing poverty data varies widely between countries, with some countries conducting surveys only every 10 years.
39. National statistical offices mainly rely on two major surveys to measure poverty: the annual EU‐SILC (European Union Statistics on Income and Living Conditions) survey that provides information on household disposable incomes and different types of material deprivations; and household budget surveys, which are typically conducted every three to five years. Some countries apply either surveys or measure income poverty from register data. See Annex II for a summary of countries’ approaches.
10
2 ConceptualBackground
2.1 Theconceptsofpoverty,inequality,andsocialexclusion
40. Poverty and social exclusion are interlinked with inequality but cannot be reduced to inequalities of income alone (Sen, 1997). Poverty is a situation in which inequalities leave some people so far away from the social mainstream that the deprivations they experience push them below what are viewed as basic standards.
41. In practice, poverty is often operationalized and measured in terms of income or consumption poverty. Poverty lines can be defined on the basis of absolute needs (e.g., the cost of a minimum food basket plus an allowance for basic non‐food basic needs), or on relative social standards that prevail in a given society at a given time.
42. One of the main sources of dissatisfaction with absolute poverty measures is that they ignore concerns about relative deprivation, shame, and social exclusion (Ravallion, 2015). Sen (1983) argued that a person’s capabilities should be seen as the absolute standard but that “... an absolute approach in the space of capabilities translates into a relative approach in the space of commodities”. Often people face interlinked deprivations (lack of education, meagre employment opportunities, etc.), which in turn reduce their income. (“When you work, you have friends. As soon as you lose your job, you have no friends at all”—UNDP Regional Bureau for Europe and the Commonwealth of Independent States, 2011, p. 8).
43. While poverty is a relatively static definition of disadvantage, social exclusion can be seen both as a process and as an outcome. As a process it pushes certain individuals to the margins of their society. It prevents their full participation in relevant social, economic, cultural, and political activities. As an outcome, social exclusion denotes the status and characteristics of the excluded individual. Examples of the many dimensions of social exclusion are: poverty, a lack of basic competencies, limited employment and educational opportunities, inadequate access to social and community networks and activities. Khan, Combaz, and Fraser (2015) provide a comprehensive overview of the topic and the related literature.
44. The social exclusion perspective evolved in in European welfare states to emphasise the denial of “social rights” (UNDP 2011). For example, Lenoir (1974) defined “the excluded” in contradiction to the ideal of citizenship and social justice. If poverty is defined in relation to income or material deprivation, social exclusion is defined in relation to such social rights as the right to work, the right to housing, the right to health services, or the right to education (Lister, 2004).
45. For Sen (2000), social exclusion means denial of freedoms. People may be unable to take advantage of an opportunity because of deliberate policies or social practices (active exclusion), or as a result of complex webs of social processes without intentions from anyone (passive exclusion). Social exclusion assigns a central role to social relational and unequal power relationships (Stewart et al., 2006). According to Silver (1995), social exclusion breaks the bond between society and the individual.
46. Its potential political implications make the measurement of poverty a necessarily delicate exercise. The vital relevance of exclusion (and its measurement) for the social fabric
can be Ireland povertyto the personsso limitStates i
47. UdefinitiobeyondThese tbelow.
Table2Compa
moneta
multidi
2.2 T
48. SievolveddistinguMore reor comb
2.2.1
49. Thinternameasurwith wothers.
7 http://e8 Most ofof best p9 Welfare“undeserconsumpwhere inphysical ecommonconsidere
illustratedjoined the
y programmadoption os, families ated as to en which the
ntil today, on of poved the meretwo aspect
2.1 ring the def
ary
mensional
Theevol
ince the 19d as a bauished by thecent devebine multip
Physical
he notion tional and res adoptedide applica
eur‐lex.europaf this section iractices in poe reform in Vicrving” poor. Tption, was distcomes were nefficiency” (Ro understandined morally jus
in the pre Europeanmes where laof a clear‐cand groups exclude theey live.” (Eu
the practierty that waly monetarts were pa
finitions of
lutiono
9th century asis for intheir focus olopments tple aspects o
subsiste
of merely national p
d to describtion by int
a.eu/legal‐conis based on thverty measurctorian EnglanTherefore secotinguished fronot “sufficientowntree, 190ng of poverty stified at that
Chapter2
ocess of En Communitaunched (Rcut definitioof personsm from theuropean Cou
ical measuras originallyry realm anrticularly e
social exclu
p
de
fpovert
different ternationalon physicalry to extenof poverty i
ence
physical spolicies fore social conternational
ntent/EN/TXThe findings of tement (2006)nd struggled tondary povertom unavoidabt to obtain the1, p.86). It ouincluded a mutime (Veit‐W
ConceptualB
11
uropean poty as one oom 1995, on of poves whose rese minimumuncil 1984)7
rement of y advocatend puts emmphasised
usion and p
static
poverty
privation
tymeas
approaches and com subsistencnd dimensionto one sin
subsistence r over 100nditions, at agencies s
T/HTML/?uri=Cthe Rio Group). to separate wty, which coulle primary poe minimum neght to be acknuch broader pilson 1986).
Background
olitical andof its poorDaly 2010)rty: “'The psources (mam acceptabl7
poverty ind by Town
mphasis on by Berghm
poverty
suremen
s to the mmparative wce, basic neons of welfagle measur
has influe0 years. 9 first withinuch as the
CELEX:31985Dp, summarized
hat was seen d result from overty which wecessaries fornowledged thpopulation, ev
d economicrest memb. Among otpoor shall aterial, cultue way of l
n Europe fonsend (1979the social
mann (1995
dynam
impoveris
social exc
nt8
measuremenwork. Theyeeds, and reare (includie.
enced scienExamples
n individual e UNDP, th
D0008&from=d and publishe
as the morallsuboptimal owas defined str the maintenahat, even in Viven if their co
integratioers in 1973her things tbe taken tural and soife in the M
ollows the 9). It pointsexclusion 5: 21) in th
mic
shment
clusion
nt of povery can be elative deprng wealth o
ntific practare the stcountries be World Ba
=EN ed in the Com
y “deserving”or even immortrictly as a situance of merelctorian Englandition was p
n. After 3, three this lead to mean cial) are Member
relative s clearly process. he table
rty have broadly rivation. or time)
tice and tatistical but later ank and
mpendium
” from the ral uation y nd the erhaps
50. Thhuman requirinroles aMoreovand alsdependnot me
51. Phis somesubject materiaare notclimatedifferen
2.2.2
52. Byinfluencconsumhousehcommueducatiagricultsubsisteaccess trequirethe menon‐ful
2.2.3
53. Todeprivarelevan“relativcontext
54. Wpast. Pe
10 An “ab“severe” “extremethe poor
he use of “needs are
ng physical as workersver, they arso expecteddent on collrely in the r
hysical neeetimes furtto change
al goods, th, after all, f, and socialnt ways.
Basicne
y the 1970ce, and wasmption needold furnituunity at laron, and cutural tools, ence conceto those facd by local ceaning of pofilment of s
Relative
ownsend (1ation. The nt to subjeve” implies t which vari
With rapid soeople living
bsolute” poverrefer to how e” poverty. In est (extreme
subsistencee mainly phenergy; th, citizens, re not simpd to act ouectively prorich industr
ds have bether qualifi because oheir relevanfixed or unvl customs. T
eds
0s, a secons supportedds of a famiure and eqrge, such altural facilitand accessept. In addicilities and scommunitieoverty to msocial roles
deprivat
1979) propterm “depctive povethat poveres over tim
ocial changg in the pres
rty line refers low the threspractice, manpoverty line).
Chapter2
e” to definehysical rathey are socparents, p
ply consumet different ovided utilitial societies
en includedied as “extof shifts in nce to the svarying. TheTherefore, m
nd formulatd strongly bily (i.e., adequipment); as safe waties. Furthes to farmingition to maservices, sues and popumaterial anspecific to e
tion
osed a thirrivation” inrty and perty standare.
e it may besent are no
to the approashold is set. Thny countries u
ConceptualB
12
e poverty haher than soial beings epartners, ners of physroles in thties and facs.
d in the cattreme” or social activsociety of the amount, qmaterial ne
tion—that by ILO. Twoequate foodand essenter, sanitatrmore, in rg. The “basaterial needch as healthulations. In d physical each individ
rd formulatncludes maerceptions ds need to
e difficult toot subject to
ach of establisherefore, not use multiple a
Background
as been critocial. Peopexpected toneighbours, ical goods eir various cilities. Thes
egorisation“severe.”10
vity and dehe day, andquality and eds turn ou
of “basic no elements d, shelter antial servicetion, publicural areas, ic needs” cds for physh care, sanithe past asneeds seemdual.
tion of the aterial and about pov
o be gauged
o justify a poo the laws,
shing a povertall absolute pbsolute pover
ticized becaple are noto perform sand frien
but producsocial asso
se needs ap
n of “absolu0 These nemand patted even the cost of foout to be soc
needs”—bewere inclund clothinges providec transportbasic needsconcept is asical survivtation, and s well as noms easier t
meaning osocial con
verty in a d against t
overty standobligations
ty line, while poverty lines nrty lines, to id
ause it imp simply orsocially demnds (Lister,ers of thosociations. Tpply univers
ute” povertyeds, howeverns. The ngoods themd depend ocially determ
gan to exeded: the mg, as well asd by and t and healts also incluan extensioal, it also education owadays, rehan to incl
of poverty: nditions whsociety. Ththeir specif
dard devises, and custo
“extreme” annecessarily refentify the poo
lies that ganisms manding , 1990). e goods They are sally and
y, which ver, are need for mselves, on work, mined in
ert wide minimum s certain for the th care, de land, n of the includes that are stricting ude the
relative hich are he term ic social
ed in the oms that
nd fer to or and
Chapter2ConceptualBackground
13
applied to the previous century. Globalisation is connecting peoples and making them more aware of differences in standards of living, while inequalities within and between countries are growing. There are, therefore, major objections to merely updating any historical benchmark of poverty on the basis, for example, of the price index.
55. Poor people are not just victims of a misallocation of resources. They rather lack, or are denied, the resources needed to fulfil social demands and observe the customs and laws of society. This realisation led to the development of “relative deprivation” approach, under which a threshold in each dimension of poverty is envisaged, according to prevailing social norms, below which withdrawal or exclusion from active membership of society is common.
56. Relative measures are most frequently used in wealthier societies. For example, the current EU definition of poverty and social exclusion combines income poverty that is defined on an annually determined relative threshold with non‐monetary deprivations that are not changed over time.11 One of the five headline targets of the Europe 2020 strategy is to reduce poverty by lifting at least 20 million people out of the risk of poverty or social exclusion by 2020. The headline indicator to monitor this poverty target is the AROPE indicator “at risk of poverty or social exclusion”, showing people who face at least one of the following conditions:
They are at risk of living in income poverty after social transfers (their equivalised disposable income is below their national at‐risk‐of‐poverty threshold, which is set at 60% of the national median equivalised disposable income);
They are severely materially deprived—they cannot afford at least four of nine items deemed to be essentials12 (a detailed list of EU deprivation indicators is available in chapter 4); or
They live in households with very low work intensity (defined as people from 0–59 years of age living in households where adults [those aged 18–59, but excluding students aged 18–24] worked less than 20% of their total potential during the previous 12 months).
57. In the European Union, people are considered to be at‐risk‐of‐poverty or social exclusion if they face at least one of these risks. Nearly 120 million people (representing about a quarter of the EU population of about 500 million) fell into this category in 2015. Around 32% of those people who were at risk of poverty or social exclusion within the EU‐28 in 2015 faced a combination of two or even all three of these risks. Figure 2.1 illustrates the combination of those risks for the EU population in 2015.
11 For further details see http://ec.europa.eu/eurostat/statistics‐explained/index.php/People_at_risk_of_ poverty _or social_exclusion. 12 The concept of “enforced lack” (wanting the item but not being able to afford it) is used. The list of deprivation items is currently under revision and a new definition is expected to consist of 12 items.
FigureNumbe2015(Millions
S
2.2.4
58. Infail to cThe popoverty
59. CoaccumucrashesconsumHousehstandarin assetborrowon. Cha
2.2.5
60. Thwhose because
2.1 ersofpers
s)
Source: EurosNote: The sufrom the tota
Wealth
n more proscapture ineossession oy conditions
ompared wulated savins in investmption overholds that ard of living tts that are
w to finance apter 3 discu
Timepov
he notion odisposable e they wo
sonsatrisk
stat (online dam of the dataal (published e
sperous ecoqualities wf wealth iss and how t
with incomeng and invetment or r time andre “asset ricthan wouldnot easily expenditurusses the is
verty
of “time poincomes m
ork long ho
Chapter2
kofpover
ata code: ilc_pa for the seveelsewhere) du
onomies evhich determs a particuo cope with
, wealth (a estments (ahousing m
d offers prch and inco be indicateconverted res, e.g., forsue of weal
verty” is bemay be higours and t
ConceptualB
14
rtyorsocia
pees01). en groups at e to rounding
ven relative mine personularly relevh them.
stock measalthough it markets). Wrotection agme poor” ced by their into moner house extelth and pov
eing increasgh enough therefore h
Background
alexclusio
risk of povertg estimates.
standards nal freedomant metric
sure) is mocan declinWealth allgainst unecan be expeincome aloy, its existeensions, moerty in furth
singly used to keep thhave relativ
Total numbeor social excl= 118.7 millio
onbytype
ty or social ex
of deprivatm and parti that pote
re stable ove dramaticlows indivixpected chcted to havne. While sence may aotor vehicleher detail.
to describehem out of vely little
r of persons atusion: on
ofrisks,EU
xclusion diffe
tion or incocipation in entially det
ver time, recally in the iduals to hanges to ve a higher mome wealthallow its owe purchases
e groups ofpoverty, btime for p
t risk of poverty
U‐28,
rs slightly
me may society. ermines
eflecting case of smooth income. material h is held wners to s, and so
f people but only personal
y
Chapter2ConceptualBackground
15
maintenance, social care, or leisure as a result.13 Joint analysis of income and time allows for the in‐depth exploration of such issues as the gender and poverty interface. It allows for a focus on vulnerable people who may be missed by traditional income poverty measures—for example, those who have to work long hours to attain incomes that are above the poverty line, or those who cannot take a job because of family care obligations, or those “time poor” individuals who could reduce their work hours without risking income poverty but keep on pushing because of stereotypes. Long hours on the job are the main cause of time poverty for both men and women—but the effects on women are more drastic. Country examples in time poverty are illustrated in Box 2.1.
Box2.1 Timepoverty—countryexamples
In Turkey, among full‐time workers, the time poverty rate of women was found to be nearly twice that of men (70% versus 37%), and among part‐time workers it was more than nine times as high (37% versus 4%—Zacharias et al., 2014). This suggests that gender differentials in time poverty do not result mainly from differences in hours of paid employment, but rather from women’s greater engagement in unpaid household production activities—leaving less time for remunerated work. Some one million non‐employed women in Turkey were estimated to be time poor because of their extensive household production engagements. The study also found a higher incidence of time poverty among the consumption‐poor as compared to non‐poor persons (65% versus 37%). This ratio is even more pronounced with regard to women: for employed men, 42% of the consumption‐poor were found to be living in time poverty, versus 29% of the consumption non‐poor. For employed women, these rates were 68% and 48%, respectively. Consumption‐poor urban and rural women had the highest rates of time poverty. Since the majority of the rural, time‐poor employed women work without pay on the family farm or enterprise, the impoverishing effects of time deficits may be harder on them than on wage workers.
Research measuring the value of unpaid care work in Albania and Serbia indicates that women perform more than twice as much unpaid labour as men (Table 1). In Albania, women spent over 5 hours per day on unpaid labour versus just less than an hour for men. In Serbia this disproportion was lower but still large, with women spending on average some 5 hours of unpaid work per day, compared to over 2 hours by men.
13 In 2013 UNECE has published the “Guidelines for Harmonizing Time‐use Surveys” to help statisticians and policymakers understand the importance of these surveys, provide guidance in their design and use, and improve the international comparability of their results.
Figure 1Time us(Minute
S
2.2.6
61. Abeen thindicatobasic nclusteredifferenmoneta
62. Foto compSwedish1971 “Lder Nat(Lenoir,conceptguide towell‐be
14 This se
15 Lenoir 16 Streete17 Counci18 Anand 19 Alkire e20 Wolff a
0
50
100
150
200
250
300
350
1 se in Albanies per day)
Source: Huma
Multidim
ll monetaryheoretically ors appear needs,16 soed disadvannt aspects oary aspects
or example,plement inch Level of LLes indicatetion” (1971, 1974) alsts are also io indicatorsing of all it
ection draws o(1974), and ten et al. (1981il of Europe (2and Sen (199et al. (2015). and De‐Shalit
Pa
ia and Serb
an Developme
mensiona
y approachechallengedin various cocial cohesintage,20 amof human wshould be m
, in the 196come measLiving Studyurs sociaux1). The aboso motivatin use; for es of “social s members
on Chapter 4 ohe history in A1), Stewart (192005, 2008). 97).
(2007).
aid work
Alban
Chapter2
bia, 2010‐20
ent Report (20
alpoverty
es lead at bd (Ringen 19conceptual fion,17 capamong otherell‐being ormeasured.
60s Europe mures (Atkiny (Johanssonx”, and Peteve mentionted the deexample, thcohesion”, s, including
of Alkire, FostAtkinson and 985).
Unpaid wo
nia
w
ConceptualB
16
011
015).
y14
best to indir988). Measuframeworksability povers. Each cor disadvanta
moved towason et al., 2n, 1973; Aller Christian ned multidievelopmenthe Council odefined as equitable a
er, Seth, SantMarlier (2010
ork
women m
Background
rect represeures of poves such as soerty,18 multoncept is dage directly
ards the de2002). Key iardt and UuLudz’s “Mamensional t of social of Europe ha“society’s aaccess to av
os, Roche and0).
Paid wo
men
entations oerty that goocial exclusitidimensiondistinct, yety, and sugge
evelopment innovationsusitalo, 197aterialien zuconcept ofindicators
as publisheability to sevailable res
d Ballon (2015
ork Unpa
Serbia
of welfare ao beyond mion and incnal povertyt each talkests that the
of social ins included t72), Jacquesum Bericht zf “social exs. Naturallyd a methodecure the loources, res
5).
aid work
nd have monetary lusion,15 y,19 and ks about ese non‐
dicators he 1968 s Delors’ zur Lage xclusion” y, other dological ng‐term pect for
Chapter2ConceptualBackground
17
human dignity with due regard for diversity, personal and collective autonomy and responsible participation” (2005, cf 2008).
63. The need to better measure non‐monetary aspects of development was reflected internationally by the Cocoyoc Declaration (1974) of UNEP/UNCTAD. The advent of the basic needs concept led to efforts to measure them using census data across Latin America beginning with Chile in 1975 (Feres and Mancero, 2001). Social indicators were also developed by the World Bank (Streeten et al., 1981). Conceptually, these efforts drew on Sen’s capability approach. Sen proposes that social arrangements should be evaluated with respect to people’s capabilities—their freedom to enjoy valuable “doings and beings”. In the 2000s, the Voices of the Poor studies (Narayan et al., 2000) provided renewed interest in wider approaches to poverty, because that is how poor people talk about it. Voices of the Poor drew on Amartya Sen’s capability approach work, and used the term “multidimensional poverty”. The MDGs (launched in 2000) also drew together existing standards in different indicators to propose a harmonised set of indicators of which monetary poverty was just one element.
64. In the 2030 Agenda for Sustainable Development, the worldwide consensus has shifted to view poverty as multidimensional. Precursors to this conceptual shift include (1) academic writings such as by Amartya Sen (1990, 1991); (2) inputs from poor persons and non‐governmental organisations, consultations leading up to the Sustainable Development Goals; (3) an increasingly visible academic literature on multidimensional poverty measurement; and (4) the pioneering leadership of countries such as Colombia, Mexico, China, South Africa, Bhutan, Pakistan, and others in using multidimensional poverty statistics to complement monetary measures and guide policy.
65. Informed by this emerging consensus, the pivotal SDG document Transforming Our World: The 2030 Agenda for Sustainable Development identifies, in its second sentence the global challenge of reducing poverty in “all its forms and dimensions,” as the foremost challenge of our time. The SDGs transparently consider poverty to be multidimensional, repeating this multiple times throughout the document. Specifically, SDG target 1.2 is, “By 2030, reduce at least by half the proportion of men, women and children of all ages living in poverty in all its dimensions according to national definitions”.
66. The SDG view of poverty as multidimensional is also shared by other organisations. For example, the World Bank’s Atkinson Commission Report “Monitoring Global Poverty” (World Bank, 2017), advised to use a six‐dimensional measure of overlapping deprivations based on the methodology of Alkire & Foster as a complementary measure of global poverty. The report advised that this measure should cover health, nutrition, education, work, living standards, and violence. Hence there is an emerging consensus that multidimensional measures of poverty should complement monetary ones, and that the methodology described in this handbook is at the forefront of this change. And since 2010, the UNDP was a pioneer in launching and reporting the global MPI and in supporting national and regional MPIs.
67. Many countries are using the MPI to measure progress towards SDG 1, and as a governance tool. An MPI supports the SDG agenda as follows:
Leave no one behind: MPI analysis tracks progress on poverty for different groups, showing who is poorest, how they are poor, and whose poverty reduces fastest.
Chapter2ConceptualBackground
18
Many MPIs are disaggregated by sub‐national regions, by rural and urban areas, and by groups such as age cohort or ethnic groups. Disaggregated analysis helps to achieve a crucial SDG aspiration: ensuring no one is left behind.
Monitor progress: the MPI is used to track and compare multidimensional poverty over time. National MPIs are used to compare regions and groups within a country;
Integrated, coordinated policy: In many countries, the MPI is used by senior policymakers to coordinate policy and to understand and track the impact of their policies on the poor, helping to break down silos and intensify policy impact.
Universal relevance: National and regional MPI measures are tailor made to context. They address poverty while reflecting policy priorities and relevant indicator definitions.
68. The theme of the 2017 United Nations High Level Political Forum for Sustainable Development was “Eradicating poverty in all its forms and dimensions”. In 2016 and 2017 many countries have been reporting progress on multidimensional poverty reduction in their Voluntary National Reviews.
2.3 Methodologicalchoices
69. The spectrum of poverty measurement approaches varies from purely monetary to non‐monetary aspects, with much variation (see Table 2.1). Measurement choices are often implicit, and can have a profound impact on results and related policies.
70. The first choice is what to measure: income, consumption, or broader capabilities? The most common approach is to measure monetary poverty, based on indicators of income or consumption as proxies for material living standards. These are the conventional poverty measures that use information on household income or consumption estimates (see chapter 3).
71. Monetary and multidimensional poverty measures are complementary. Both are valuable for identifying poor people and shaping policy. They provide different insights. For instance, the Multidimensional Poverty Index (MPI) developed at Oxford University with the UNDP’s Human Development Report Office uses ten indicators to measure three critical dimensions of poverty at the individual level: education, health, and material living standards (For more on this see Chapter 5.) These indicators measure deprivations in health and educational outcomes as well as in access to key services such as water, sanitation and electricity. In the mid‐2000s, the number of people living in extreme poverty in Europe and Central Asia was 12 million according to the MPI, while 23 million lived on less than PPP$1.25/day. Nonetheless, multidimensional poverty was relatively low in most of these countries.
72. What can be surprising is the common finding that people who are multidimensionally poor, or deprived in non‐monetary indicators, are not necessarily income poor. Divergences between monetary poverty and multidimensional poverty indicators mean that neither is a sufficient proxy for the other; both need to be measured (see Box 3.5). Moreover, reducing non‐monetary deprivations often requires different policies than reducing income poverty.
Chapter2ConceptualBackground
19
73. A helpful survey of empirical research on commonly observed mismatches between different poverty indicators is found in Nolan and Whelan’s 2011 book Poverty and Deprivation in Europe. Nolan and Whelan offer a systematic review of “why and how non‐monetary indicators of deprivation can play a significant role in complementing (not replacing) income in order to capture the reality of poverty in Europe” (p. 1). Another literature survey is found in Chapter 4 of Alkire et al. (2015).
Table2.2Differentapproachestopovertymeasurement
Unidim
ensional
Monetary
Income based
Absolute poverty lines
National thresholds specific for individual countries, in the national currency
1. Cost of basic needs
2. Subsistence minimum
Internationally comparable thresholds
3. Severely poor with income below 1.9 PPP$
4. “Just poor” with income below 3.1 PPP$
Relative poverty lines
Share of the median (or mean) income
5. Relative low income (example: below 50% or 60% of the contemporary median equivalised income in each country)
6. Relative low income anchored at a fixed point in time
7. Weakly relative poverty line
Expen
diture based
Absolute poverty lines
National thresholds specific for individual countries, in national currency
8. Cost of basic needs
9. Subsistence minimum
Internationally comparable thresholds
10. Severely poor with expenditures below PPP$1.90/day
11. “Just poor” with expenditures below PPP$3.10/day
Relative poverty lines
Share of the median (or mean) expenditure
12. Relative low expenditure (example: below 50% or 60% of the current median equivalised expenditure in each country)
13. Relative low expenditure anchored at a fixed point in time
14. Weakly relative poverty line
Food energy intake (FEI) 15. Nationally specific FEI‐based poverty rates (varies by climate conditions, rural/ urban distribution, type of occupation, etc.)
Multidim
ensional
Deprivations
16. Indicator dashboards
17. Indices of multiple deprivation, including material deprivation
Multidimensional poverty estimates – internationally comparable (following the methodology developed by OPHI and used for international comparisons and in the Global HDRs published by UNDP)
18. Multidimensional poverty index (thresholds for the various dimensions)
Official national multidimensional poverty indices, following the methodology developed by OPHI
19. Severely poor
20. Moderately poor
Source: Modified from Ivanov and Kagin (2014).
74. Thmost mWhile examplwould lSection
2.4 M
2.4.1
75. Powhen mcategorframe sreach gpeople health refugeeway. Thsamplininstanceetc.) costandar
2.4.2
76. DThis is disaggrconcernemploy
77. Thone hamisuseddiscrim
2.4.3
78. WSuch a shared World
he second fmonetary anthere are es of its meead to a fe 6.8).
Measure
Non‐cove
overty statmeasuring ries of peopsince they groups, suchwho are ininstitutionses in campshese groupsng concernse, while chovered, theyrd surveys.
Disaggre
isaggregatioa key aspeegation enning age, seyment and h
he collectioand, respond. On the ination and
Equivale
When monescale is comconsumptiBank, 2000
fundamentand non‐monno interna
easurementw robust in
ementis
erage
istics shoulpoverty thple who maydo not liveh as homelen institutions; Roma; pes or illegal s usually res, or becaushildren in iny may lack t
egation
on is necesect of Agenntails surveex, educatiohealth statu
on of data tndents mayother han
d unequal tr
encescale
tary measummonly useon, housing0). Econom
Chapter2
al question netary approationally at and deternternationa
ssues
d in theoryrough povey be likely te in househess people ns, includingeople in temimmigrantsquire specise they maynstitutions mthe social s
sary to provda 2030’s ey design ton level, occus, and ethn
that are disy view thisnd, policymreatment by
es
ures are useed to adjustg and spec
mies of scal
ConceptualB
20
is who shouoaches), or greed meaminants exlly compara
y cover all erty surveyto be poor aholds. This (including sg elderly camporary acs are notoral approachy face specmay have tkills needed
vide a detaaspirations to allow focupation, anicity, can a
saggregateds as sensitimakers needy ethnicity.
ed, the chot householdcific needs e arise, fo
Background
uld identifypeople theasures on ist. This Guable indicat
of the popys it shouldare frequenis of particstreet childre homes, commodatriously diffiches—eitherial forms oftheir basic d for inclusi
iled picture“to leave
or the collnd place oflso be of ke
d by ethnicive informad statistica
ice of equivd resources(Lanjouw, r example,
inadequacyemselves (susubjective ide proposeors on subj
pulation of d be recogtly omittedular concerren), drug children’s hion or hostcult to accer because of deprivationeeds (fooion—which
e of certain no one belection andf residence.ey importan
ty can be cation that al informat
valence scas in order toMilanovic a, by sharin
y: the analyubjective popoverty,
es future wective pove
interest. Hgnised that from the srn for the users, sex whomes, andtels; prisoneess in a sysof unrepreseon or exclusd, clothes, is not capt
population hind”. Mosd analysis . Disaggregance.
challenging.can potention to rect
ale can be do take into and Stefanog expendit
yst (as in overty)? national
work that erty (see
owever, certain ampling hard‐to‐workers, d mental ers; and stematic entative sion. For shelter, tured by
groups. st often, of data ation by
. On the tially be tify any
decisive. account o, 1998; ures on
housinghousehattain aand the
2.4.4
79. Was last yis impoNetherlincreaselower eduratiobe paidincomeproportthe high
80. Inand tramuch oamong remain consecuperiod. vast magroup o
81. Kunderstare alwprospecborrowlongitudevents standarcan hel
2.4.5
82. DdifferenMDG re
g, utilities, old membea certain lee elderly are
Poverty
When analysyear? Have ortant to mlands in thes in labouend of the in. In additid to the ext mobility otion of peoher the repo
n the Repubansient comof 90% of thouseholdspoor in th
utive years,Only 14% oajority of thof househol
nowing thtanding theways unwelcts if indiv
w. However,dinal data leading indrds. Longitup policymak
Reportin
iscrepanciences in the eporting co
car or newers may alsvel of well‐e particularl
dynamic
sing povertythey just fameasure pohe 1980s aur market pncome diston to the mtent to whior volatilityople experieorted pover
blic of Moldmponents rthe people s in terms ohe next pe, the analysof househohe populatids remaine
e length oe short‐ andcome, theyviduals and these tactone can unividuals intudinal poverkers adopt
ngonpov
es betweenways in whntext, for e
Chapter2
wspapers. Ao influence‐being. Mely affected
cs
y trends, it allen in poveoverty fromand 1990s,participationtribution shmagnitude ach poverty y and the encing poverty rate will
ova during revealed thclassified a
of rank, a veriod. Usingsis showed olds were noon was expd poor thro
of time thd long‐term y may not household
tics are unlinderstand to and out orty analysisbetter safet
vertyand
n internatihich the assxample, de
ConceptualB
21
Apart frome the amouasures of tby the choic
is importanerty, or is thm a longitu high leven did not rowed that and duratiois recurrenshorter therty at leasl be.
1997‐2002,at poverty as poor (Beery large frag the set othat aroundot poor in aposed to pooughout the
hat a houimpact of pthreaten s
ds can redikely to be the processof poverty, can also idty nets or o
dinequal
onal and sociated indspite the ex
Background
m householdnt of incomhe incidencce of equiva
nt to ask: Arheir povertyudinal perspls of econeduce povepoverty spn of low innt (Fourage he durationt once dur
, a decompwas main
eegle, 2004action of thef panel hod 25% of hoany of the foverty durine entire per
usehold hapoverty. Altsubsistenceuce expensufficient ises behind and the assdentify waysther inclusi
lity
national ddicators arexistence of
d size, the me or consuce of poveralence scale
re the poor y recurrentpective. Foomic growerty. Howevells were gecome statuand Layte,
n of poverting the rep
osition of ply chronic, 4). That is, e poor in anuseholds inouseholds wfour surveyng 1997‐20iod.
s been pothough shoe or significditure, runn the long cross‐sectisociated ims in and ouon policies.
databases oe defined anthe official
age or geumption nerty among e.
the same t? In other wor example,wth and sigver, data frenerally of s, attention, 2005). Thety, the higported perio
poverty intoaccountingdespite trany year are nterviewed were poor y rounds. W02, a sizeab
oor is crurt spells of cantly damn down savrun. Only bonal statist
mpact on thet of poverty.
often resund reportedglobal list o
ender of eeded to children
this year words, it , in the gnificant rom the short in n should e higher gher the od, thus
o chronic g for as ansitions likely to in four in every
While the ble core
ucial for poverty
mage life vings or by using tics: the eir living y, which
ult from d. In the of goals,
Chapter2ConceptualBackground
22
targets, and indicators issued by the United Nations,21 most countries provided data on only some of these (see Table 2.2).
Table2.2 ReportingonMDGindicatorsininternationalandnationaldatabases
Country
1.1 Proportion of population below
PPP$1/day
1.1a Proportion of population below the
national poverty line
1.2 Poverty gap ratio
1.6 Proportion of employed
people living below
PPP$1/day
Albania I B2 B*
Armenia B2 B2 B* B*
Azerbaijan I B2 B* N*
Belarus I B2 I
Bosnia and Herzegovina I B2 I
Bulgaria I B B*
Croatia I B I
Czechia I N2 I
Georgia B2 B B2*
Hungary I B2 I
Kazakhstan I B2 B* I
Kyrgyzstan B2³ B2 I
Latvia I B2 I
Montenegro I B B*
Republic of Moldova I B2 B*
Romania I B2 B*
Serbia I B2 I N*
Slovakia I N2 I
Slovenia I N I
Tajikistan B²,³ B2 I I
TFYR Macedonia I B B*
Turkey B B2 B* I
Turkmenistan I N2 I
Ukraine I B2 I N*
Uzbekistan I N I
Source: United Nations Economic Commission for Europe, Report on the differences between national and international reporting about MDG 1 prepared for the Regional Workshop on poverty and employment indicators of the Millennium Development Goal 1, Almaty, 27‐28 September 2011.
Note: I – International database, N – National database; B – Both databases; B2, N2 – At least two definitions are used for this indicator in the national database; * ‐ In the national data series, the index is computed on the basis of the national poverty line; ** ‐ In the national data series, the index is computed for children under three years of age; ¹ ‐ The international series also presents data disaggregated by gender; ² ‐ The international series also presents data disaggregated by gender; ³ ‐ The national series also presents data disaggregated between rural and urban areas. Empty cells indicate that the corresponding figures do not appear in either database, for any of the years taken into consideration (from 1990 to 2009).
83. Some poverty indicators may appear only in international data series, whereas others may only appear in national series. Moreover, for indicators that are included in both sets of
21 See http://unstats.un.org/unsd/mdg/Host.aspx?Content=Indicators/OfficialList.htm
Chapter2ConceptualBackground
23
databases, definitional variations can produce significant differences in the values reported for national and international purposes (see Table 2.3).
Table2.3 Sharesofpopulationlivingbelowincomepovertylines,forselectedcountries,2009‐2012
Country Population below national
poverty line, per cent
Population below PPP$1.25/day, per cent
Year
Albania 14.3 0.5 2012
Armenia 32.4 1.8 2012
Belarus 7.3 0.0 2011
Bulgaria 21.3 1.9 2011
Czechia 9.7 0.0 2011
Estonia 18.7 1.0 2011
Georgia 14.8 14.1 2012
Hungary 14.1 0.1 2011
Kazakhstan 6.5 0.1 2010
Kyrgyzstan 36.8 5.1 2011
Latvia 19.3 1.1 2011
Lithuania 18.6 0.8 2011
Montenegro 9.4 0.2 2011
Poland 17.1 0.0 2011
Republic of Moldova 17.5 0.2 2011
Romania 22.7 0.0 2011
Russian Federation 13.0 0.0 2009
Serbia 24.7 0.1 2010
Slovakia 13.3 0.3 2011
Tajikistan 47.2 6.5 2009
Ukraine 8.9 0.0 2010 Source: The official United Nations site for the Millennium Development Goals Indicators, maintained
by the United Nations Statistics Division. Available from http://mdgs.un.org/unsd/mdg. Last accessed 20.12.2016.
Note: 0.0 refer to the percentage smaller than 0.05. “Population below national poverty line” refers to the percentage of the population living below the national poverty line, which is the level deemed most appropriate for the country by its authorities. Population below PPP$1.25/day refer to the percentage of the population living below the international poverty line of PPP$1.25/day.
84. Complications in measuring inequality result from the fact that the most common international databases that show income distribution data for the countries of the region —such as POVCALNET or SWIID— often present data that differ from what can be found on the public websites of the national statistical offices in the region.
2.4.6
85. Mpovertydashboseparatinsightsdifferen
86. Thindividucomposmultidimpovertydeprivainclude
Measurin
Multiple appy. Broadly ards of caretely and us into the dnt datasets.
he second gual deprivasite measumensional y thresholdations, and income or
ngnon‐m
proaches haspeaking, tefully definenidimensioifferent asp
group consiation indicaures (Alkirepoverty, ths) is usualloften uses expenditur
Chapter2
monetary
ve emergedthese can ed and validnally. Takepects of pov
ists of Multators that e and Fosthe identificaly based ona countingre poverty a
ConceptualB
24
ypoverty
d in responsbe divided dated sociaen togetheverty consid
idimensioncontain d
ter, 2011; ation of whn the joint g approach among the d
Background
y
se to the neinto two l indicatorsr, these mdered one b
al Poverty Ieprivation see also ho is poor distributio(Atkinson, dimensions
eed to measgroups. Th, which pre
measures caby one; the
Indices (MPthresholdsChapter 5(according
on of indivi2003). Thes.
sure non‐mhe first consent each inan offer ey can also d
PIs), which cs into aggr). In the to one or dual or hose may or m
monetary nsists of ndicator mpirical draw on
combine regated, case of several
ousehold may not
3 M
3.1 C
3.1.1
87. Aused infor low
88. Inover timas privbenefitscomponHandbocontextdeem ithe goo
89. Cowants, for by alevels oof curreimportaO’Dea (the distgroundpoverty
90. Mrelativeprovideabsolutbeen esmany cthresho
91. ByaverageincomepovertyparticipCouncilresourcaccepta
Moneta
Concepts
Introduc
s set out instrument foliving stand
ncome refeme. It includvate penss/social tranents of hoook on Hot as it allowmportant. Tods and serv
onsumptionwhilst consa householdof consumpent economant determ(2012) and tribution os. Howevery measures.
Monetary poe. Absolute e a given mte measure stablished bcountries. Fold, which s
y contrast, e situation w or expendy that movpate fully in in 1975, wces are so inable in the
aryPov
sandm
ction
the previoor measurindards.
rs to the odes wages,ions, invesansfers. Theousehold inusehold Inws people tThose withvices they n
n is the useumption exd. Considertion or con
mic well‐beinminant of ecothers (seef consumpr, there are . The pros a
overty meapoverty linminimum sis the PPP$by the WorFor exampletood at $12
relative mwithin a socditure. The ves beyond society. Anwhich statenadequate asociety in w
verty
methods
us chapter,ng poverty.
ngoing flow salaries, anstments ae main intcome in mcome Statito satisfy thh low incomneed to part
e of goods axpenditure ed simply, nsumption eng. Many economic we Noll, 2007tion rathera number oand cons of
asures can es representandard of$1.90/day (rld Bank. Hoe, the Unite2,071 a year
measures utciety. Typicarationale absolute n example s, “People as to precluwhich they
25
, monetary Income or
w of economnd money eand other ernational icro statististics (UNECheir needs mes typicallticipate fully
and serviceis the valueand everytexpenditureconomists w
well‐being th7 for a revier than incoof reasons weach appro
broadly bent the valuef well‐being(in 2011 priowever, diffed States Cr in 2014 fo
tilise poverally, these lfor such andestitution of such a dare said tode them frolive.” This d
indicators aconsumpti
mic resourcearned thronon‐govestandards ics are contCE, 2011). and pursuey have resty in the soc
s to directlye of consumhing else bee can be rewould arguhan incomeew) argue thme, on bowhy many coach are dis
e divided ie of a set leg. Perhaps ices) line foferent absoCensus Burer a single ad
rty lines thines are basn approachto consid
definition is be living iom having adefinition is
are by far thon are usua
ces that a hough self‐emrnment sodescribing tained in thIncome is e many othtricted capaiety in whic
y satisfy a pmption goodeing equal, garded as hue that conse alone. Inhat it is preoth theoretcountries prcussed late
nto two tyevel of resothe most
or extreme olute poverteau uses andult househ
at are set sed on eithh comes froering indivthat set oun poverty ifa standard os operation
he most coally used as
household mploymentources, anthe conce
he Canberraimportant
her goals thacities to cch they live.
person’s neds and servipeople withaving lowesumption isndeed, Brewferable to cical and prrefer incomr in this cha
ypes: absolources necewidely recpoverty, whty lines aren absolute hold.
in relationer mean orom a definiduals’ caput by the Euif their incoof living connalised thro
mmonly s proxies
receives t as well nd cash pts and a Group t in this hat they onsume .
eeds and ces paid th lower er levels s a more wer and consider ragmatic me based apter.
ute and essary to cognised hich has used in poverty
n to the median nition of acity to uropean ome and nsidered ough the
Europewith eqsimilar medianis to beamong of meddisabilit
92. Dindicatonot necenjoy hhigher iin moneof conselderly levels dprivate educati
93. Sudescribexamplincome(Depart
3.1.2
Recommobserva
94. If to collevery dihousehpaymenSimilarlwhole h
95. Thcase ofresourcparentsconside
96. Thissues, equitab
an Commisquivalised dapproach i
n equivalisede used to mcountries”)ian incometies.
espite theirors of povecessarily imphigher livingincome in tetary poversumption mpeople witdespite adehouseholdon, healthc
uch limitatied in publice, the Unit” in their ptment for W
Unitofo
mendation ation should
data are coect detailedfficult or iold or famnts may bey, it is chahousehold t
he need to f families wces of theirs. The meaering the ho
he measurehowever.
bly amongs
sion’s indicdisposable iin their stad disposablmonitor pro) focuses one—using da
r usefulnesrty. Importaply low stang standardshe future). rty indicatomay in parth physical lquate finand resourcescare, water
ons are oftcations bothted Kingdompublished sWork and Pe
observati
1: In prodd be the ho
ollected thrd data frommpossible
mily unit. Foe allocated allenging toto its individ
measure inwith childrr own and surement oousehold as
ement of ecFirst, it is t all memb
Chapter
ator based ncomes betistics, sette income oogress towan reducing ta that are
s and ubiquantly, low hndards of liv by using sAdditionallrs. Similarly reflect indlimitations tncial resours do not neand infrast
en recognish by nationm Departmstatistics, wensions, 201
on
ducing datausehold, fo
rough housm all membeto allocateor exampleat the hou
o allocate edual membe
ncome at thren. The chrely predo
of such intthe basic st
conomic resgenerally
bers of the
r3Monetary
26
on the propelow 60% oting relativeof the entireard meetingthe share odisaggrega
uity, there ahousehold iving. A houavings or tay, levels of y, and depedividual chothat limit thces). More ecessarily rructure.
sed in the wal governm
ment for Wowhilst Euros16; Eurostat
a on incomr both prac
sehold surveers of the he economice, certain tusehold or expenditureers.
he househohildren wilominantly ora‐househotatistical un
sources at tnecessary
e household
yPoverty
portion of inf the natioe poverty te populationg SDG10—“of national ated by age
are a numbncomes or sehold withaking on dewealth areending on thoices or noheir mobilitgenerally, reflect acce
way that moments and inork and Petat reportst, 2016).
me or consctical and co
eys, it is ofthousehold. c flows to types of infamily, rats that are
old level is pl typically on intra‐hoold transfernit, the need
the househoto assumed. In realit
ndividuals lnal medianhresholds an. SDG indic“reduce inepopulation
e, sex, and
ber of limitalow levels h a low incoebt (based typically nohe thresholon‐monetarty may havemonetary mess to basi
onetary povnternationansions refe on “at‐risk
umption, tonceptual re
ten impractMore imposingle indicome fromher than thcarried ou
perhaps beshave few,
ousehold trrs is, at besd to do so is
old level pree that resoy, there m
iving in houn. The OECDat 50%‐60%cator 10.2.1equality witns living belpersons liv
ations to mof consumpome may beon expectaot taken in lds used, loy constraine low consumeasures bc services
verty indical organisatiers to “relatk‐of‐povert
he normal easons.
tical and exortantly, it viduals wit
m social prohe individut on behalf
st illustrateif any, ecansfers frost, difficult,s removed.
esents a nuources are
may be an
useholds D uses a % of the 1 (which thin and ow 50% ing with
monetary ption do e able to ations of account w levels nts (e.g., umption based on such as
ators are ons. For tive low y rates”
unit of
xpensive is often thin the otection al level. f of the
ed in the conomic om their , but by
mber of shared
unequal
distributhe houtime (JPonthieand its collectioas the b
97. Inmeet bclearly consideBank’s accountis likelyhousehapproacwomancalled e
3.1.3
98. Adoes noseek tohouseh
Recommindicatohouseh
3.1.4
99. Th
Either (separatof the hpersonsthemserelated pool the
100. ThConferePopulatbenchm
ution of resousehold. Thenkins, 19eux, 2013) himplicationon challengbasic unit of
n determiniasic needs needs to ber householPPP$1.90/dt for econoy to need aold, but nch also assn is the samequivalence
Unitofa
lthough incot change to improve told.
mendation ors describolds below
Househo
he Canberra
(a) a personte room (or housing unis who comelves with fopersons oneir income.
his definitience of Eution and Hmark for pov
ources betwhe limitatio91) and rehas attempns for poveges have limf account re
ng whetheror allow pabe taken ind income oday and PPmies of scaa higher incnot necessaumes that
me as that ne scales are
nalysis
come and cohe fact thatthe welfare
2: Povertying, for exthe poverty
olddefini
a Handbook
n living alonrooms) of ait to form pmbine to ocood and posnly or of unr
on is baseropean StaHousing (Uverty measu
Chapter
ween men ans of this aesearch on ted to betterty statistimited progremains inte
r a given levarticipation nto accountr consumptP$3.10/dayle within hocome to enarily three the level oneeded by often used.
onsumptiont poverty ise of individ
y statistics sxample, they line.
ition
k (p. 64) set
ne in a sepa housing upart of a muccupy the ssibly other related pers
ed on the atisticians (NECE, 2006urement.
r3Monetary
27
and womenassumptiontime poveer understaics. Howeveress in this agral to almo
vel of housein society,
t. The simption per capy poverty liouseholds. njoy the samtimes the
of resourcesan 8‐year‐o. These are
n are norma experiencedual citizen
should be re number
ts defines a
parate housnit but doesulti‐person hwhole or pessentials fons or of a
definitionCES) Recom6) and sho
yPoverty
n or betwee have beenerty (and iand intra‐hoer, substanarea. As a rost all publi
ehold econthe numbe
plest approapita. This is tnes. HowevFor examplme standare income. s needed bold boy. To discussed la
ally measured by indivins, regardle
eported at of individu
household
sing unit or s not join whousehold opart of a for living. Thcombinatio
of a privmmendatioould be co
n different n widely rein other arousehold shntial methoresult, reliaished pover
omic resouer of peoplach to deathe methodver, such ane, a househrd of living Additionall
by, for examaccount foater in this
ed at the hoduals. Policess of their
the individuals in a p
as:
r who occupwith any of tor (b) a grohousing un
The group mon of both. T
vate housens for the onsidered t
generationecognised foreas, for eharing of redological ance on hourty statistics
rces is suffie in the holing with thd used in thn approachhold of threas a singlely, the permple, a 40‐yor these pochapter.
ousehold lecies should r status wit
dual level, wopulation l
pies, as a lothe other ocoup of two nit and to may be compThe group m
ehold used 2010 Censthe recomm
ns within or some example, esources nd data useholds s.
icient to ousehold his is to e World h fails to ee adults e person r capita year‐old ints, so‐
evel, this likewise thin the
with the living in
odger, a ccupants or more provide posed of may also
in the suses of mended
101. Inthe bascases. Fduring home dbetweetime is legal adshould the allo
102. Abasis, itchallengwho arearrangearrangeinclude
103. Infor the with ot
3.1.5
104. PoHowevenot alwmeasurgroups.
3.1.5.1
105. Cofor shelto be agroup ogreat mcategorassistedinstitut
106. Thcommucollectiohousehcarried respond
n line with tsis for houseFor examplthe weekenduring termen multiple spent). In ddress shouinclude witocation of in
lthough it it is recognige in both e not usual ement. Theements on an examina
ndividuals acompilatioher populat
Populati
overty stater, as with ways straigrement, as .
1 Comm
ommunal elter and sub legal bodyof persons. majority ofries: resided‐living faciions; religio
he vast majunal establion, thoughold incomeout for thding countr
the CES cenehold meme, those wnds (place om‐time (plaresidencesthe case ofuld be usedthin the mendividuals.
s recommeised that dmeasuremeresidents oe Australianhousehold ation of equ
and familiesn of poverttion sub‐gro
ioncover
istics shouall social stghtforward it is often t
munalesta
establishmebsistence ary for the puInstitutionsf institutioences for silities and wous instituti
jority of houshments, lh there are or consumhe latest edries’ incom
Chapter
nsus recombership. Thho work awof usual resce of usuas (place of uf a child trd. In all casetadata the
ended that ifferent fament and conof a househon Bureau oeconomic uivalisation
s not living ty statisticsoups that ar
rage
ld cover alatistics, theor even
the case th
ablishme
ents or instire being prourpose of los usually hanal househstudents; hwelfare insons; and wo
usehold staargely duere additionmption in sudition of the micro‐sta
r3Monetary
28
mendationshe recommeway from fasidence is fl residenceusual resideuly spendinses, those idefinition
householdsmily structunceptual teold but do lof Statisticsresources a practices.
in private hs and these re sometim
l of the poe practical lpossible. T
hat poverty
nts
itutional hoovided by aong‐term inave commoholds are hospitals, ctitutions; morker dorm
atistics colle to the prnal challenguch establishe Canberratistics cov
yPoverty
s, the placeendations damily homefamily homee is family ence shouldng equal timnvolved in of househo
s should beures and inderms, in parive there sos plan to eand housing
householdsare discusses omitted
opulation oimitations oThis is a is more pr
ouseholds cn institutionhabitation on facilities consideredconvalescenmilitary baritories.
ected througractical diffges associashments. Tha Handbooered comm
e of usual redistinguish ae during the), school chome), or d be the adme in multithe measu
old used an
defined ondividual chrticular the ometimes uexamine thg in future
s provide a sed in the from officia
or sub‐popuof data collparticular revalent am
comprise pen. An instituand provisishared by td to fall int and oldracks; corre
gh social suficulties assated with he survey ook revealedmunal estab
esidence sha number ofe week andchildren awa child alteddress wheiple addresurement of d the appro
n a usual reoices do primpact of
under a shahe impact research. T
practical chnext sectioal statistics.
ulation of iection meaissue for
mong hard‐t
ersons whoution is undion of servithe occupanto the fod people’s ectional an
urveys do nosociated withe defin
of country pd that noneblishments
hould be f special d return way from ernating ere most sses, the poverty oach for
esidence rovide a children red care of such This will
hallenge n, along .
interest. an this is poverty to‐reach
se need derstood ces to a nts. The ollowing homes;
nd penal
ot cover ith data ition of practices e of the such as
Chapter3MonetaryPoverty
29
university halls of residence or institutions for long‐term care. The collection of consumption data from communal establishments is likely to be particularly challenging, both conceptually and in practice.
3.1.5.2 Homeless
107. Those with no usual place of residence are also not covered by standard household surveys designed to measure income or consumption. However, they also typically represent some of the poorest and most vulnerable individuals in society. Homeless households include those living in temporary or insecure accommodation, as well as those who are sleeping rough.
108. Whilst it may not be possible to include homeless households within standard household surveys, it is important to consider alternative ways in which such households can be captured in information about poverty. The approach used is likely to vary across countries according to the information available. In Nordic countries, for example, data on population registers may be of some use. In Australia, the wider scope and coverage of the Census of Population and Housing, which uses a special enumeration strategy to target rough sleepers, people in supported accommodation and those living in overcrowded households, allows the limitations of private household survey collection to be overcome. Elsewhere, it may be possible to make use of information collected by local government or other agencies, as well as the voluntary sector.
109. Italy’s experience of collecting data for the homeless population is described in Box 3.1.
Box3.1 CollectingdataonthehomelessinItaly
Surveys of homeless people, conducted in 2011 and 2014, were part of a research project on the conditions of people living in extreme poverty, following an agreement between Istat, the Ministry of Education and Social Policy, the Italian Federation of Associations for the Homeless (fio.PSD), and the Italian Caritas organisation.
The first stage of this research was to define the reference population (Istat, 2014) using the common operational definition for comparative research provided by the European Typology on Homelessness and Housing Exclusion (ETHOS) (Amore et al., 2011). The definition adopted was based on the roofless and houseless subgroups. The former includes those living rough and those staying in a night shelter. The latter includes people living in homeless hostels and other forms of temporary accommodation. Not included in the study were those defined as experiencing housing exclusion, rather than homelessness, under ETHOS. This group includes those living temporarily with family or friends, or those living in extremely overcrowded conditions.
Sample design for homeless people requires a time‐location sampling type, where the individuals belonging to the population of interest are identified by the places they frequent and their periods of frequency. For this study, the places that homeless people frequent are taken as the locations providing services to meet their needs, as well as the public spaces in which they
Chapter3MonetaryPoverty
30
spend time (De Vitiis et al., 2014).
In contrast to household surveys, no pre‐existing sampling framework for the referenced population was present. Two approaches were instead considered: the first involved selecting individuals at canteens and night shelters (known to be frequented by large numbers of homeless people); the second focused on public spaces. Both approaches had limitations related to incomplete coverage of homeless populations, and to the risk of including people multiple times.
In night shelters and canteens, the potential multiple counting issue results from possible repeated use of the same services; this can be addressed through the identification of people surveyed. Addressing this challenge is more complex for a survey conducted in public spaces, however. Full population coverage is not guaranteed with either approach: whereas the first method will not capture those homeless people who are not using either night shelters or canteen services, the second method (outdoor sampling) is unable to guarantee full coverage of the territory.
The first approach was adopted, as capturing all the required information in centres providing canteen and night shelter services seemed more feasible. The sample base was therefore constructed by referencing service providers (i.e., indirect sampling methodology).
In the first survey, conducted in 2011, the list of services was constructed in two phases, prior to the survey of the homeless: (i) a census of the organisations offering services to the homeless in 158 Italian municipalities was conducted; and (ii) an in‐depth survey of services provided was also conducted (Istat, 2011, 2013). The survey of the homeless, which was the third phase of the process, was conducted over a period of thirty days, in order to include a larger number of service users.
The sample design randomly distributed the interviews over the opening hours and days of the centres in the reference month, and included all the centres involved in the two previous phases. A two‐phase sample plan was used: the first stage involved selecting the survey days, while the second focused on services provided.
The number of homeless people was estimated by measuring the number of links between each interviewed individual and the services used in the week immediately preceding the interview: this was done by keeping individual weekly diaries recording individual visits to the various centres on the reference list. In this way, the estimates were not distorted by double counting.
The diary was filled by the homeless people who were able to respond to the interview, with the help of interviewers and service operators, given the habitual behaviour that generally characterises the homeless population. A simplified diary version has been submitted to people using only canteens or only night‐time accommodation services. Imputations procedures have been adopted for diaries partially filled or not filled at all. In the first case, an intra‐record probabilistic imputation was made, whereas for total non‐responses, both the weekly link number and the individual weight were imputed, taking into account the centers characteristics and their non‐response rates.
In the 2014 survey, it was estimated that 50,724 homeless people (in the months of November and December) used at least one canteen or night shelter in the 158 Italian municipalities. This corresponds to 2.43 homeless people per thousand of the population regularly registered with the municipalities covered by the survey—a higher value than in 2011, when it was 2.31 per
Chapter3MonetaryPoverty
31
thousand (47,648 persons). It should be noted, however, that only 69% of the homeless people surveyed stated they were entered in the civil registry of an Italian municipality. This is figure stood at 48% among foreign nationals and 97% among Italians.
The duration of the condition of homelessness had also increased in comparison with 2011: the share of those who had been homeless for less than three months had declined from 29% to 17%, while the shares of those who had been homeless for more than two years increased from 27% to 41%, and of those who had been homeless for more than four years from 16% to 21%. Some of the characteristics of the homeless population in 2014 are highlighted in the Figure 1 below.
Figure 1 The main characteristics of homeless people in Italy, 2014
3.1.5.3 RomaandTravellerpopulations
110. The Roma and Traveller populations are often underrepresented in poverty indicators and social statistics more broadly. This can be for a number of reasons, including, for example, unauthorised and some authorised caravan sites not being represented on the sampling frameworks used for income and consumption surveys.
111. Targeted surveys can help to better understand poverty amongst these groups. This is the approach used by UNDP, UNICEF, the World Bank, and the EU Agency for Fundamental Rights, who have collected data on poverty and social exclusion through surveys of Roma populations in the Balkans and EU countries respectively (see e.g., UNDP, 2009; UNICEF, 2010, 2011, 2012, 2014a, 2014b; European Union Agency for Fundamental Rights [FRA], 2014). This has allowed for comparisons of levels of monetary poverty and material deprivation of the Roma with the rest of the population in those countries (see Box 3.2).
Box3.2 UNDP’sexperiencecollectingRomapovertydata
UNDP has addressed Roma inclusion issues by running specialized surveys to collect comparable and trustworthy information about Roma poverty and living conditions.
The first such survey was conducted in 2002 in five Central and Southeast European countries, and has since been repeated several times and greater numbers of countries. Data
producepovertyEurope membe
A speciwith Umonito
The promonito
3.1.5.4
112. Infor houincludeunder‐cwho arein flats/
113. Opracticaflats/apbarriersfinancia
114. Hsurveyspovertyon the tpoverty
RecommhousehThis mafor peobig datusers ab
3.1.6
115. Inindicatoand genindicato
ed by the my picture fro(Ivanov, Kl
er states, pr
al samplingNDP providring living c
ovision of ring system
4 Other
n addition tusehold surd within thcoverage ine not fluent/apartment
Other reasonal obstaclepartments, s, and unwal informati
igh incomes. Although y statistics, type of threy (inequality
mendation olds. It is reay involve rople who ara), in ordebout the co
Disaggre
n addition tors that arender. Such ors for the
most recenom a humaing, and Kagroviding com
g methodoloding recomconditions a
comparablm of the Dec
rdifficult‐
to these groveys in moe surveyed official stat in the mas or in hous
ns why soms as difficbeing presewillingness ton is collect
e and wealtinformatiotheir abseneshold usedy) where th
3: It is recoecommendresearch ame homelessr to estimaoverage of t
egation
o indicatorse disaggregadisaggregatoverall pop
Chapter
nt survey con developmgin, 2012). mparable da
ogy has beemmendationand discrimi
e survey‐bcade of Rom
‐to‐reach
oups (who ost countried populationatistics. Thein languageses of multi
me sub‐popuulties in aent at the ato participated).
thy individuon about sunce may affed. And it cae threshold
ognised thated that NSImong comms (as in Box ate povertyhe publishe
s for the ovated by suction is highpulation ma
r3Monetary
32
onducted inment perspeThis surveyata.
en developns for usingination (Ško
ased poverma Inclusion
populatio
are typicalles), there ans, are ofteese include es of the cople occupat
ulations areaccessing inaddress whate in offic
uals are alsch individuect who is in have a did is set with
t the majorIs explore tmunal estab3.1), or ut
y in difficuled poverty s
verall populh dimensiohly desirableay mask su
yPoverty
n 2011 proective for cy was also c
ed to addreg ethnicity obla, Leonči
rty data wn (http://ww
ons
ly absent frre also groen very diffifragile and
ountry in whtions.
e harder to ndividual ahen interviecial surveys
so often diuals does noidentified arect impactrespect to
ity of poverthe feasibililishments, ilising altert‐to‐reach statistics is
lation, it is ns of socio‐e for a numbstantial va
vided a coountries in onducted b
ess the speas a stati
ikas, and Ště
was crucial ww.romade
rom the samups who, wicult to reac disjointed hich they liv
cover in suddresses fewers makes (particula
ifficult to rot feature das being in pt on measuthe mean.
rty statisticsty of extenincluding hnative datapopulation essential.
important t‐economic vmber of reasariations in
mprehensivthe Eastern
by EU‐FRA f
ecifics of thiistical indicěpánková, 2
for settingcade.org/).
mpling framwhile in thech, leading householdve, and tho
urveys inclufor those le contact, larly when p
each in hodirectly in apoverty, derements of
s only coverding this coostels and a sources (ingroups. In
to produce vulnerabilitsons. First, the povert
ve Roma n part of or 11 EU
is group, cator for 2009).
g up the
meworks eory are to their s, those se living
ude such iving in anguage personal
ousehold absolute pending relative
r private overage. shelters ncluding nforming
poverty y as age poverty ty levels
Chapter3MonetaryPoverty
33
experienced by different population subgroups. Understanding which groups experience the highest levels of poverty is important for targeting policy interventions effectively. For example, high levels of poverty among retired people will likely require different policies than those targeted at reducing poverty among children.
116. Monitoring disaggregated indicators can help ensure that no‐one is left behind as countries make progress towards reducing poverty. Some research on MDG achievement found that most progress in poverty reduction was made amongst those easiest to reach, or in situations easiest to address—leaving many of the poorest and most vulnerable behind (e.g., Save the Children, 2010).
117. The production of such disaggregated data can provide challenges for compilers of poverty statistics. For one thing, producing disaggregated data requires larger survey samples, and may also require more complex sample designs (especially where certain sub‐groups make up a small proportion of the overall population). All of this can add to the costs of data collection.
118. Some disaggregations need to be interpreted with caution as the needs of different subgroups may not be fully reflected in standard poverty lines. For example, the material needs of people with disabilities are often greater, due to both additional costs as a result of goods and services needed due their disability as well as higher costs for some other items compared with individuals without disabilities. Some studies have attempted to account for this using a variety of methods (see, for example, MacInnes et al., 2014).
119. In addition to this recommended minimum disaggregation, further disaggregation would in many areas be of policy relevance. For example, the child population could be disaggregated into smaller age groups as there are often significant differences in poverty rates between these age groups. Different rationale can be applied to this, for example related to policy objectives (for example pre‐school; school age; secondary school); or age groups (0‐4; 5‐9; 10‐14; and 15‐17). Disaggregations by other variables, including ethnicity and occupational group may also be relevant in some countries.
Recommendation 4: Given their importance, poverty data should be disaggregated whenever possible. As a minimum, poverty indicators for the UNECE region should be disaggregated by age, sex, employment status, household type, disability status22 and urban/rural population.
It is further recommended that the following classifications be used for these breakdowns.
Age:
0‐17 (children)
18‐24
25‐49
50‐64
65 and over
22 Where possible, disability status should be measured using the short set of disability questions developed by the Washington Group http://www.washingtongroup‐disability.com/washington‐group‐question‐sets/short‐set‐of‐disability‐questions/
Employ
EmURO
Househ
O
Tw
Tw
Tw
O
O
Urban/r
Pr In Pr
3.2 W
3.2.1
120. ThConfereHandbo
23 Whereincome/c24 This claof existinavailable
yment statu
mployed nemployedetired
Other outsid
hold type:
One‐person
wo adult ho
wo adult ho
wo adult ho
One adult ho
Other.
/rural24:
redominantntermediateredominant
Welfare
Incomec
he conceptences of Laook, is as fo
Householdservices) householdother such
Householdthe net wother fina
Householdand self‐e
e possible, emconsumption assification shng classificatio from OECD (2
us23:
d
e the labou
households
ousehold w
ousehold w
ousehold w
ouseholds w
tly urban ree region tly rural reg
measur
concepts
tual definitabour Statisllows:
d income cothat are rd at annuah irregular a
d income reworth of theancial or non
d income memploymen
ployment statreference perhould be basedon methodolo2011) and Eur
Chapter
ur force
s;
ithout child
ith one chil
ith two or m
with children
egion
gion
res
sanddefi
tion of hosticians (ILO
onsists of areceived byl or more fand typicall
eceipts are ae householdn‐financial a
may be definnt); (ii) prop
tus should relriod. d upon the pogies used for rostat (2015b
r3Monetary
34
dren;
d under 18;
more childre
n under 18;
initions
usehold inO, 2004) an
all receipts y the housefrequent inly one‐time
available fod through aassets or an
ned to coverperty incom
late to the (m
opulation deninternational ).
yPoverty
;
en under 18
;
come estand adopted
whether moehold or bytervals, bureceipts.
or current coa reduction n increase in
r: (i) incomeme; (iii) inc
ost frequent)
nsity of local acomparison w
8;
blished byin the 201
onetary or y individuat excludes
onsumptionof its cash,n its liabiliti
e from empcome from
status during
reas rather thwithin the UN
y the Inter11 Canberra
in kind (goal memberswindfall ga
n and do no, the disposes.
loyment (bothe produ
g the
han self‐reporECE region ar
national a Group
ods and s of the ains and
t reduce sal of its
oth paid ction of
t. Details re
Chapter3MonetaryPoverty
35
household services for own consumption; and (iv) current transfers received (other than social transfers in kind); and (v) social transfers in kind.
121. This definition shows what, in principle, should be included in a comprehensive measure of household income (see Table 3.1). In practice, income definitions adopted by individual countries may be more limited in scope, as some elements of household income may not be collected or modelled (this is typically the case, for instance, with unpaid domestic services, with services provided household consumer durables, and by social transfers received in kind).
122. The component elements of income can be aggregated in order to produce selected measures for particular analytical and policy purposes. The sum of income from employment (section 1 in Table 3.1) and income from household production of services for own consumption (section 3) is referred to as income from production. Adding income from production to property income (section 2) gives primary income. Total income is the sum of primary income and current transfers received (section 4); from this measure it is possible to obtain disposable income, which is total income less current transfers paid (section 8). Total and disposable incomes are the most commonly used income aggregates.
Recommendation 5: It is recommended that annual (equivalised) disposable income be the main income measure used for poverty measurement, as this reflects the actual income that individuals within a household have available for spending or saving. However, to provide additional insights into the nature of poverty in a country or area, compilers of poverty statistics may also wish to make use of supplementary income measures, such as income before social transfers.
Disposable income should be defined in line with the practical definition set out in the Canberra Group Handbook (2011), with the exception of the net value of owner‐occupied housing services, which should be excluded (see Section 3.2.8.3).
123. The relationship between those identified as at‐risk‐of‐poverty based on disposable income and those who are dependent on social transfers in Germany is explored in Box 3.3.
Chapter3MonetaryPoverty
36
Table3.1 IncomecomponentsintheconceptualdefinitionoftheCanberraGroupHandbook
Canberra 2011 conceptual definition
1 Income from employment
1a Employee income Wages and salaries Cash bonuses and gratuities Commissions and tips Directors’ fees Profit‐sharing bonuses and other forms of profit‐related pay Shares offered as part of employee remuneration Free or subsidised goods and services from an employer Severance and termination pay Employers’ social insurance contributions
1b Income from self‐employment Profit/loss from unincorporated enterprise Goods and services produced for barter, less cost of inputs Goods produced for own consumption, less cost of inputs
2 Property income 2a Income from financial assets, net of expenses 2b Income from non‐financial assets, net of expenses 2c Royalties
3 Income from household production of services for own consumption 3a Net value of owner‐occupied housing services 3b Value of unpaid domestic services 3c Value of services from household consumer durables
4 Current transfers received 4a Social security pensions/schemes 4b Pensions and other insurance benefits 4c Social assistance benefits (excluding social transfers in kind) 4d Current transfers from non‐profit institutions 4e Current transfers from other households (cash only)
5 Income from production (sum of 1 and 3)
6 Primary income (sum of 2 and 5)
7 Total income (sum of 4 and 6)
8 Current transfers paid
8a Direct taxes (net of refunds) 8b Compulsory fees and fines 8c Current inter‐household transfers paid 8d Employee and employers’ social insurance contributions 8e Current transfers to non‐profit institutions
9 Disposable income (7 less 8)
10 Social transfers in kind (STIK) received
11 Adjusted disposable income (9 plus 10)
Source: UNECE, 2011.
Chapter3MonetaryPoverty
37
Box3.3 Personsatriskofpovertyandbeneficiariesofsocialtransfers:Differentconcepts‐differentpeople?AcasestudyfromGermany2014
The German system of social reporting in official statistics (“amtliche Sozialbericht‐erstattung”) provides a wide range of comparable national and regional (“Länder”) level data. One data source looks at the beneficiaries of the social security system. Another source provides data on relative poverty (at‐risk‐of‐poverty rate). Statistics drawing on both sources are published online.
Poverty at the national level can be measured via EU‐SILC (Statistics on Income and Living Conditions) methodology, which covers individual and household living conditions in both monetary and non‐monetary terms. One of the key indicators is the at‐risk‐of‐poverty rate, based on household disposable income (after social transfers).
For analysis at the regional level, the utility of SILC is limited by the fact that its sample size is currently 0.03% of the German population. Therefore, the at‐risk‐of‐poverty rate on NUTS 1 (“Länder”) and NUTS 2 levels (plus additional regional breakdowns) is not based on SILC data, but rather on information produced by the “Mikrozensus (labour force survey)”—an annual household survey that samples 1.0% of the total population. The at‐risk‐of‐poverty rate only reflects current income; situational needs, wealth status, and actual housing costs are not considered.
In contrast to the at‐risk‐of‐poverty rate, the number of persons relying on social transfers is a different concept describing people who are dependent on social assistance in order to get by. In Germany, the most claimed social assistance for this group is the so‐called unemployment benefit II (based on Book II of the Social Code, known as the “Hartz IV” Act). All people who are able to work, but are unemployed and in need (and who are not entitled to unemployment insurance under Social Code III) receive transfers for themselves and (if applicable) their dependents. This includes assistance for costs of accommodation as well as mandatory health insurance. Similar transfers are provided for persons unable to work and persons above retirement age in accordance with Social Code XII.
Data on social transfers usually take the form of administrative data and are available for recipients of different types of social status as well as at various regional levels. In contrast to the household survey Mikrozensus, administrative data are available at the NUTS 3 level (“Kreise”) and beyond. Therefore, administrative data are the main source for studies on inequality and poverty at the municipal level.
Although these two indicators (the poverty rate and the number of recipients of social transfers) are derived from different data sources, are based on different definitions of poverty, and are available at different spatial levels of aggregation, they are both widely accepted and used in various studies on social development. Often they both complement one another. However, they also may lead to different results and conclusions about who is at risk of poverty.
The Mikrozensus is able to apply both concepts simultaneously to the same person, allowing the degree of overlapping between the two measures to be examined. Analysis of the Mikrozensus data for the year 2014 shows that:
t
w
sd
ts
3.2.2
124. Thmeasurbe influrelating
3.2.2.1
125. Inperspecthey dedesirabare avasaving.
126. Dmakersprotect
127. Dincome
17.9% of ththe at‐risk‐population
Of this gropoverty threceiving tr
More than were not rpeople are not fulfil thlevel of wesome may do not reinformation
14% of the their incomsometimes particularlymember nepoverty thr
Advanta
here is no re is preferauenced by bg to the use
1 Advan
ncome mective, incomeem importble as a welfailable to a
Direct policy due to thetion paymen
Disaggregati by source
he populatio‐of‐poverty can be con
oup, just unreshold—aransfers to a
half of the receiving soconsideredhe conditioalth or the fulfil the coport this tn or fear of
potentiallyme after tthe case w
y high. Addeeds can alreshold.
gesandd
simple ansable for meboth conceof income
ntages
asures home allows ptant to theifare measuan individua
y link. Inco direct policnts to famil
ion by incom (wages, p
Chapter
on (14.2 mithreshold sidered as p
nder one tnd was reavoid pover
potentially ocial transfd at‐risk‐of‐pons to recelow housin
onditions toto the socbecoming s
poor weretransfers, wwhen transfeitional earnlso push in
disadvan
swer to theeasuring moeptual and are set out
ouseholds’ eople to sar lives. In pre as, in geal or house
ome basedcy levers thies below th
me source censions, so
r3Monetary
38
llion peopleand/or re
potentially
third (32.6%eceiving sorty, they are
poor (53.4%ers. One expoverty witive social ang costs/reno receive social securitstigmatized
receiving twhich was ers to covernings and acome after
ntagesof
e question onetary povpragmatic ibelow.
command atisfy their nparticular, aeneral, it is ehold for e
poverty mat exist throhe poverty
componentocial protec
yPoverty
e) in privateeceive sociapoor.
%) had an cial transfee still consid
%) were at xplanation th regard toassistance (nt). Anotheocial assistay authoriti).
ransfers buabove the
r costs of acallowances r transfer p
incomea
of whetheverty. In prissues. Som
over resneeds and an indicatoran effectiveeither cons
measures aough, for exline.
ts. In generaction receip
e householdal transfers
income beers. In othdered at risk
risk of monmay be th
o their curre(for exampler potential ance but fories (possib
ut were not e poverty ccommodatto meet adpayments ab
asapove
er income oractice, the me of the m
ources. Frpursue manr such as die proxy for umption (if
are often axample, the
al, it is posspts, intra‐ho
ds either livs. This part
elow the ather words, k.
etary poverhat, althougent income,le because explanatior some reasly due to
at risk accothreshold.
tion and headditional hobove the at
ertymeas
or another decision w
main pros a
rom a conny other gosposable inthe resourf they so w
appealing toe targeting o
sible to breaousehold tr
ve below t of the
t‐risk‐of‐despite
rty—and gh those , they do of their
on is that son they lack of
ording to This is ating are ousehold t‐risk‐of‐
sure
welfare will likely nd cons
nceptual oals that ncome is rces that wish) or
o policy of social
ak down ransfers,
Chapter3MonetaryPoverty
39
etc.) when analysing poverty. This provides advantages both in terms of understanding poverty within a certain group, and as a quality check for the data, via possible comparisons with other sources.
128. Ease/cost effectiveness of measurement. In general, data on household income are relatively cost effective to collect, compared with consumption expenditure. Even if administrative data are not available, the relatively small number of potential sources of income means that data collection can potentially be relatively straightforward. This makes income‐based poverty measures particularly useful when either the cost of collecting consumption data would be prohibitive, or where precision at the national or regional level (via surveys based on larger sample sizes) is a priority.
3.2.2.2 Disadvantages
129. Links between income and living standards are not always clear. Income is a measure of potential rather than achieved living standards. As a result, current income may either overstate living standards (when the family is saving, as not all the income translates into current consumption) or understate them (when current consumption is not constrained by income, through dissaving or borrowing) (Atkinson et al., 2002).
130. Sensitivity to short‐term income fluctuations for some groups. Linked to the above point, incomes for some population groups may be particularly susceptible to short‐term fluctuations, which are typically not reflected in achieved living standards. These groups include the self‐employed, agricultural workers, and the temporarily unemployed.
131. Some components are difficult to measure. While data on some income components such as wages and salaries are relatively straightforward to collect, other components (including self‐employment, and especially including agricultural work) are much more difficult to accurately measure, largely because of the difficulty in separating out business costs and revenues. In developing countries, income data may be particularly difficult to collect, and their accuracy difficult to verify, because most of the population may be engaged in the informal sector. There is evidence of increasing imputation rates (due to refusal or inability to reveal specific income components) over time, in recent years (see Meyer et al., 2015).
132. Evidence of under‐reporting. Evidence from a range of countries suggests a general tendency for income to be under‐reported by low‐income households (e.g., Meyer and Sullivan, 2011; Brewer and O’Dea, 2012). There are a number of possible reasons for this. In part, people may forget income they have received during the reference period from sources such as intra‐household transfers, social transfers, or home‐produced items they have sold. Second, people may be reluctant to disclose the full extent of their income, for privacy reasons—particularly if any of that income has either not been disclosed to the tax authorities or has been obtained through illegal activities (e.g., Deaton and Grosh, 2000). Underreporting of income can be minimised to a certain extent through interviewer training in probing and explaining confidentiality rules, asking survey respondents to refer to payslips and other documentation where possible, and the use of data from other sources such as administrative data.
3.2.3
133. Insurveysexamplinforma
134. Borecommhybrid asourcesinformaof CES adminis
135. Hsourcesper capmeasur
136. Th(UNECE
Box3.4Theco
Many Epurposerespondsocial seducati
Two quadminisproblemensure
The Itamain inand Dolinkageto each
The ma
Datasou
n most cous developede, the Noration on the
oth types omended thaapproachess (such as taation not avcountries;
strative dat
ousehold is used for npita measurring poverty
he collectioE, 2011).
4 ombinedus
EU countriees, driven mdents (Jäntstatistics areion records
uality dimenstrative datms due to linternal co
lian SILC (ITncome compnatiello, 20. The individh person by
ain steps in t
urcesfor
untries, hod specificalrdic countre distributio
of source hat producerss, taking infax records ovailable from Box 3.4 a in Italy.
ncome datnational accres, withouty.
on of incom
seofsurve
es are conmainly by ttti, Törmälee populatio.
nsions shouta: timelineate data densistency.
T‐SILC) has ponents, be13). Surveydual matchthe Italian t
the IT‐SILC
Chapter
househo
usehold inly for that ies), registon of house
ave their os of povertyformation oor benefits dm registers.provides a
ta are alsoounts prodt distributio
me data is
eyandadm
sidering inhe need toehto and Mon registers
ld be carefuess and coelivery by d
developed eginning in 2y data and aing‐key is ttax authorit
production
r3Monetary
40
oldincom
come micrpurpose. Hers (adminhold incom
own strengty statistics uon some comdata), and m. This approan example
o available uction are tonal inform
covered in
ministrativ
creased uso reduce theMarlier, 201s, tax regist
ully looked omparabilitydata owners
a multi‐sou2004 and dadministratihe tax codety.
process are
yPoverty
me
rodata primHowever, innistrative dae.
ths and limusing incommponents omatching thoach is beine of the c
from natiotypically onation. As su
more deta
vedata
ses of adme cost of da13). The maters, social
at when coy. Using res, and to ex
urce data ceveloping five records e that is the
e set out in
marily comen a numbeata) are th
itations. Wme as a welfof income fhis with survg taken in iombined u
onal accounly availableuch, they ar
ail in the C
inistrative ata collectiain adminissecurity da
onsidering tegisters canxtensive pr
ollection mfurther in laare combine personal I
Figure 1 be
e from hoer of counthe main so
Where possibfare measurrom adminvey data concreasing nuse of surv
nts. Howeve as aggregare of limite
anberra Ha
data for ston and burstrative souata, and he
he increasen cause timactices inte
methodologyter years (Cned via exacD number a
elow.
ousehold ries (for ource of
ble, it is re adopt istrative ntaining numbers vey and
ver, the ates and d use in
andbook
tatistical rdens on urces for alth and
ed use of meliness ended to
y for the Consolini ct record assigned
Figure 1Main st
The intdata, th
Comparadminis
Total ho
The disincomeshows right odistribu0.33) (D
Table 1Indicato
Year
Year
Gini
These dcomponrespondin the s
1 teps in the
egration prhe exact ma
rison of sustrative dat
ousehold n
stribution oe, imputed higher leveof the whoution is less Delle Fratte
ors on inco
rly mean (eu
rly median (
coefficient
differences nents in thdents forgeurvey of we
IT‐SILC prod
rocess showatching proc
urvey data a can impro
et income
of total hourent, the
els (mean aole income concentratand Laricci
me levels a
uros)
(euros)
are partly ehe adminisetting to meell‐to‐do sec
Chapter
duction pro
wn in this dcedure, and
with the ove the qua
usehold neinterest rend median distributio
ted, accordia, 2015).
and distribu
IT‐
explained bystrative daention themctions of ho
r3Monetary
41
ocess
diagram cond detection
final integality of the f
et income (payments values inc
on curve (sing to the G
ution by dat
‐SILC survey
23,454
19,676
0.33
y the preseta but not
m. Also, the ouseholds
yPoverty
nsists of theand reconc
grated datafinal income
(the sum oon mortgarease by arsee Figure Gini coefficie
ta source, 2
y
nce of sevet in the spresented d
e acquisitioiliation of d
a highlightse estimates
of total disge) for theround 50%)2). In addent value (0
2011
IT‐SILC
35
29
0
eral secondaurvey datadata show t
on of admindiscrepancie
s how linki.
posable hoe IT‐SILC e) and a shifdition, the 0.30 compa
integrated
5,036
9,611
0.30
ary and lowa, possibly the under‐c
nistrative es.
ing with
ousehold stimates ft to the income red with
w‐income due to
coverage
Chapter3MonetaryPoverty
42
Figure 2 Income distribution by data source, 2011
The majority of changes from the original survey data lead to increases in reported incomes, which sometimes significantly change the position of individual households within the income distribution: 8.3% of households belonging to the bottom two quintiles according to the survey data are, according to the IT‐SILC integrated data, actually in the top two quintiles.
At‐risk‐of‐poverty rate
The overall at risk of poverty rate in 2011 was estimated at 21.2% using survey data and 19.6% when using the linked data (IT‐SILC). Overall, 87% of people were classified in the same way by both the survey and IT‐SILC data.
Table 2 People at risk of poverty by data source, 2011
IT‐SILC survey data IT‐SILC integrated Percentage
At risk of poverty At risk of poverty 13.8
At risk of poverty Not at risk of poverty 7.4
Not at risk of poverty At risk of poverty 5.8
Not at risk of poverty Not at risk of poverty 73.0
Total 100.0
Analysis of the data from these two sources, as well as examining administrative data on its own, reveals that the integrated use of sample and administrative data improves the quality of final estimates on income variables, in comparison with the estimates obtained by
0
5
10
15
20
25
30
4 11 18 25 33 40 47 55 62 69 76 84 91 98 105 113 120 127 135 142 149 156
Annual income, thousands of euro
Percentage
of
households
‐> IT‐SILC integrated
‐> IT‐SILC survey
conside
Even if povertyas a codata co
Some idata (fosurvey and to p
The chothe onereconciadminison the o
3.2.4
Recommon cons
137. AConsumof consneeds. T
138. If people value of
139. Exhousehhousehthe pay
140. Dexpendacquire
ering only su
the differey, it is still pmplement, ollection.
ncome comor exampledata qualitypursue mor
oice is a trae hand, vele survey strative datother.
Consump
mendation sumption ex
ccording tomption, andsumption gThese good
through the
as in‐kind goods and
investment
from the ho
as transfers
the goods themselvesf the goods
xcluded froold needs, olds. Theseyment of int
ifferent thiture. With ed or taken
urvey data o
ent sourcespreferable toften throu
mponents c, pension iy (taking intre detailed a
de‐off betwrsus possiband adminta are collec
ptionexp
6: Where cxpenditure.
o the OECd Wealth (Ooods and sds and servi
e purchase
income frod services
ts (e.g., port
ousehold’s
s in kind fro
and services or were res and service
om consumpsuch as cu
e are classifterest on co
heoretical the acquisipossession
Chapter
or administ
show subso make useugh the com
an be satisncome) andto account analysis.
ween higheble reductioistrative dacted accord
penditur
consumptio.
D FramewOECD, 2013services usece are obta
of goods an
om employeproduced
tion of crop
own produc
om other ho
es included eceived fromes should b
ption expenurrent tranfied as non‐onsumer cre
approachesition approan of, regard
r3Monetary
43
trative data
stantial cone of registermbination o
sfactorily esd administrmissing‐dat
r accuracy aons in dataata at the ding to adm
e:concep
on is used a
ork for Staa), consumed or paid ined:
nd services
ers, from sby the ho
p provided b
ction of goo
ouseholds o
in consumm elsewhere included
nditure are sfers to go‐consumptiedit.
s exist foach, goods,dless of whe
yPoverty
.
nsistency inrs not as a sof multiple
stimated exrative data ta mechanis
and coverag timelinessmicro levinistrative,
ptsandd
as a welfare
atistics on ption expefor by a h
in the mark
self‐employousehold),
by share‐far
ods and serv
or from busi
ption expenre as incomas consump
expenses novernment, on current
or measuri and serviceether they
n terms of isubstitute fdata sourc
xclusively ucan be usesms and me
ge of the tas (due to thel) and corather than
definition
e measure,
the Distrinditure repousehold t
ket;
yment (throor from p
rming tenan
vices; or
nesses.”
nditure werme in kind, tption expen
not directlysocial orgaexpenditur
ing househes are incluhave been
identifying for survey des and mul
using admined both to easurement
arget populhe time nensistency (n statistical
ns
it should b
bution of presents “tho directly m
ough the bproperty o
nt);
re producedhe notionalnditure.
y aimed at anisations ore. Also exc
hold consuuded when tpaid for or
those in data, but ti‐modal
nistrative improve t errors),
ation on eeded to (because criteria)
be based
Income, he value meet its
arter of or other
d by the l market
meeting or other cluded is
umption they are r; in the
case of and serpaid foFramewfor mosapproac
141. Thvehiclesexpecteconsumcan be provide
142. Indurablethan th
3.2.5
3.2.5.1
143. Bethat ultBecauseachieve
144. Shfor examore v1957), wexpectasubject fluctuatfor exaagricultproduct
145. Dsuggestquality of the althougreferenquestio(thoughday exp
f goods, regrvices are ior. While thwork recomst consumpches.)
he assumpts, electricaled to be usmer durableviewed as es a flow of
n the OECD es and owne initial pur
Consump
1 Advan
etter proxytimately sate of this, ced living sta
hort‐term fmple, adjusolatile, a finwhich suggations ratheto seasontions. Nevemple, whentural societtion cycles.
Data qualityted that, at of consumpincome disgh this depece period
ons about coh see belowpenditures c
gardless of ncluded whhe use app
mmends useption goods
tion does nl appliancesable for ms, it does noa producingservices to
Frameworer‐occupiedrchase of th
ptionexp
ntages
y of materiatisfies a houconsumptiondards than
fluctuationssting savingnding that gests that der than theial fluctuatiertheless, exn householties, where
y is arguableast in couption expenstribution (eends on thused. This onsumptionw for exceptcan lead to
Chapter
whether thhen they arproach may of the acqs and servic
ot hold fors, furniture
more than aot normallyg entity thaitself as the
k, it is therd housing the capital ite
penditur
al living stausehold’s neon expenditn income.
s are smoogs or drawiled to Frieddecisions mir current inons, but thxpenditure ds make pu incomes a
bly better, auntries suchnditure datae.g., Meyere collectionpotential
n are usuallytions). Also,fewer recal
r3Monetary
44
hey have bere used, regy more clouisition appces, there w
r dwellings e, clothing, year). Why consume tat invests ine consumin
refore the fhat is incluems.
e:advan
andards. It ieeds and wture can be
thed out. Hng on wealdman’s “peade by conncome. Thishese are suvolatility murchases inand expend
at least at h as the Una is better tr and Sullivn method, advantage y seen as le using a sholl errors.
yPoverty
een used. Wgardless of sely equateproach for pwill be little
or for consand the liken a housethem immedn those itemg entity.
flow of servded as con
tagesand
s the consuwants withine considere
Households lth and borrmanent innsumers ares does not mupposedly smay be hig large voluditures are
the bottonited Statesthan that fovan, 2011; as well as is common
ess sensitiveorter refere
With the uswhen theye to consupractical ree difference
sumer durake (which wehold purchdiately. Ratms as capit
vices obtainsumption e
ddisadva
umption of n the houseed a more
can smootrrowing. Incncome hypoe based onmean that csmaller thah under somes and lo highly cor
m of distris and the Unor income toBrewer ansuch factonly ascribee than questence period
se approachy were acqmption, thasons. (In pe between
bles such awould normhases a dwether, the hoal expendit
ned from coexpenditure
antages
goods and ehold budgedirect mea
th consumpcomes may othesis” (Fr long‐term consumption seasonal ome circumow frequencrrelated wi
ibution. It nited Kingdowards the d O’Dea, 2rs as the led to the fations aboutat least for
h, goods uired or e OECD practice, the two
s motor mally be elling or ousehold ture and
onsumer e, rather
services et limits. asure of
ption by, also be iedman, income
on is not income stances, cy, or in th farm
is often dom, the bottom
2012) —ength of act that t income r day‐to‐
3.2.5.2
146. Uundereexpendmay beshows tremainefood awSullivan
147. Irmay, insuch asmay noWhile, afor indexpend
148. Inare largsteps toacquirepovertyin the r
149. Dof the mexpensihousehapproxipolicy m
150. Inindividulimitatiodespiteconsumthreshowill play
3.2.6
3.2.6.1
151. Dthroughincome
2 Disad
nder‐reporstimated initures for ge seen as sothat while red stable ovway from hn, 2015).
rregular exp the majori food, drinkot be properat the aggreividual houiture not ne
ndirect policgely an issuo either ince those resoy (by, for execorded co
Data collectimethod useive—considold budgetimately oncmonitoring i
ndividual cual choices ons, such a adequate
mption expeold is at or cy a significa
Datasou
1 House
ata on conh househol), or more
dvantages
rting of certn surveys goods and seocially unacreporting raver time, thhome, shoe
penditures oty of cases,k and transprly reflectivegate level, usehold it ecessarily in
cy levers. Bue of persocrease the rources, or xample, by nsumption
ion compleed, the collederably mort surveys isce every fiin those cou
hoice. In sor non‐moas lack of financial
enditure‐basclose to subant factor in
urcesfor
eholdsur
nsumption d budget se general
Chapter
s
tain expendbecause ofervices thatceptable (eates for somhere have bs and cloth
on high val, provide a gport, expenve even whethis may prwill lead tndicative of
Beyond subsonal choiceresources aotherwise improving expenditur
x/expensivection of dere so than s such that,ive years. Tuntries.
some circumonetary conmobility, mresources. sed measurbsistence mn consumpti
consump
rveys
expendituresurveys expsurveys co
r3Monetary
45
ditures. Cerf under‐rept are illegal e.g., alcoholme of the bbeen noticeahing, and al
ue items. Wgood indicaditure on hen a survey rovide a relto ‘noisy’ f material liv
sistence mi. As a consavailable forimprove thsocial housre of househ
ve. As highlietailed houincome da, in many EThis makes
mstances, nstraints. Fomay have Whether ires dependsminimum levion expendi
ptionexp
e are colleplicitly focusovering a w
yPoverty
rtain groupsporting by (e.g., illega, gambling)iggest comable declinecoholic bev
Whilst a diaation of typhigh value itwith an aniable measudata with ving standa
nima, levelsequence, r consumpthe materiasing), these holds.
ighted in thusehold conata. In partiEuropean cs them uns
low levels or example,low levels ndividual cs on the levvels, it is obiture aroun
penditure
cted throused on expwider rang
s of expendresponden
al drugs, pro). In additioponents of es for someverages (se
ry period ofical expendtems is usuanual refereure of houslevels of rds.
s of consumwhile govetion, increal living stanmay not b
he previous nsumption eicular, the countries, thsuitable for
of consum, elderly peof consumchoice presvel of threshbviously lesd that poin
e
gh househopenditure (age of topic
ditures are tnts. These ostitution) oon, recent econsumpti
e categoriese Bee, Mey
f a couple oditure on caally infrequeence period sehold expetotal consu
mption expeernments mse the potendards of te directly r
section, regexpenditurecost of conhey are cor effective
mption mayeople with mption expesents an ishold used. Wss likely that.
old surveysas well as cs. While d
typically include
or which evidence on have s such as yer, and
of weeks tegories ent, and is used.
enditure, umption
enditure may take ential to those in eflected
gardless e data is nducting nducted poverty
y reflect physical enditure ssue for Where a t choice
s, either possibly data on
consumaggrega
152. Hdesignethe prodetailedindividuunderst
153. Byby the informain conscoveredmany Eindicato
3.2.6.2
154. Dbasis. Dcompleintervieearlier electron
155. Boexpensereferen
156. BoRetrospbeing realso plaseveral
Recommfor a gapproac
For inteincomecountri
3.2.7
157. Loexpend
mption expeate nature o
ousehold bed primarilyoduction of d expenditual and htanding of p
y contrast, World Ba
ation necessiderably led include heEuropean coor aids the d
2 Retro
ata on consData collectted by the ews or by tperiod. Datnic) comple
oth methodes when usce period (w
oth retrosppective inteeported, foace large bumonths (al
mendation given popuches.
ernational c be the mes as well a
Usingmu
ooking at thiture, and
enditure foof the data
budget survey to providethe nationure categohousehold poverty with
the Living Snk in the sary to meaess detail tealth, educountries, thdevelopmen
ospectivea
sumption eted retrosperespondenelephone. a collected eted by the
ds imply rissing retrospwhen there
pective inteerviews canr large houurdens on rethough usu
7: Where ulation, the
comparisonain welfares increasing
ultiplew
he intersecmaterial de
Chapter
r householmake them
eys can vare expenditural accountsories, whichcharacteri
hin a count
Standards M1980s, areasure living than a typation, emphis wider rant of effect
andongo
xpenditure ectively arent. InterviewRetrospection an ongohousehold.
sk of errorspective mete is telescop
erviews andn be very tseholds or espondentsually the rec
both incomere is valu
s of poverte measure, g availability
welfarem
tion of mueprivation (
r3Monetary
46
ds are avaim unsuitable
ry widely in re weights fs. Consequeh limits thestics and ry.
Measuremee multi‐purstandards t
pical houseloyment, mange of daive policies
oingcollec
are collectee collected w data are ive collectiooing basis a
s: omission thods) or thping of expe
d diaries cime‐consumfor househos, as reporticording peri
me and consue in utilis
ty across thgiven its wy in other a
measures
ltiple indicaanother fo
yPoverty
ilable from e for poverty
terms of dfor measureently, theseeir potentiother ite
ent Study (Lrpose housthrough hoehold budgmigration, anta collectedfor poverty
ction
ed retrospeby an intergenerally con means tare collected
of certain he inclusionenditures an
an pose laming when olds with coing periods iod is daily f
sumption esing povert
e UNECE rewidespreadareas of the
ators for thrm of pove
national ay measurem
esign; howees of inflatioe surveys foal to colleems releva
LSMS) survesehold survusehold coet survey).nd savings. d alongsidey reduction.
ectively andrviewer or vcollected ththat data ad with a dia
expenses (en of expendnd no bound
arge burdenmany expeomplex strucan vary frfor one to t
expenditurety measure
egion, it is r usage amregion.
he same peoerty covered
ccounts, agment.
ever, almoson and to feocus extensect informaant to in
eys, first deveys collectnsumption . AdditionaAs with EU
e the main .
d/or on an ovia a questihrough facere collectedary (either p
especially fditures outsding intervi
ns for housenditure iteuctures. Diarom severalwo weeks).
e data are aes based o
recommendong EU an
ople using d below) m
gain the
st all are eed into sively on ation on ncreased
veloped ting the (though al topics U‐SILC in welfare
on‐going ionnaire e‐to‐face d for an paper or
or small side the ew).
seholds. ems are aries can days to .
available on both
ded that d OECD
income, measures
Chapter3MonetaryPoverty
47
can produce new insights. Where a household is income poor but is maintaining expenditure and is not materially deprived (those in income poverty only), this may indicate that the household is able to draw on savings or access loans either informally or formally to maintain living standards. In some cases, such behaviour may be driven by the knowledge or expectation that household income will increase in the near future—for example, for those starting a new job soon or students. However, many households of this type are likely to remain vulnerable to poverty.
158. Expenditure poverty in the absence of either income poverty or material deprivation can be seen as an indicator of uncertainty over future income levels and a lack of accumulated wealth or assets which could be used to maintain living standards if income drops. This may occur in employment that has no guaranteed future income (for example, for those in short‐term employment and the self‐employed), or on so‐called “zero‐hours” contracts (Serafino and Tonkin, 2017a).
159. Often it is not possible to examine these different poverty measures with the same dataset. However, techniques such as statistical matching open up the possibility of using synthetic datasets (see Box 3.5). Box 3.5 describes work on the statistical matching of data for measuring poverty undertaken in the EU. Another example from the US of fusing datasets is provided by Garner & Gudrais (2017).
Box3.5 Comparingpovertyestimatesusingincome,expenditure,andmaterialdeprivationdata
As part of the Eurostat funded Second Network for the Analysis of EU‐SILC (Net‐SILC 2), Serafino and Tonkin (2017b) carried out a study comparing people’s exposure to poverty in a range of countries (Belgium, Germany, Spain, Austria, Finland, and the United Kingdom) using data for income, expenditure, and material deprivation. The concept of material deprivation is introduced in more detail in Chapter 4 (section 4.4).
As there is currently no single data source providing joint information on all of these variables for households or individuals, it was necessary to first statistically match expenditure data from the 2010 round of the household budget survey (HBS) with income and material deprivation data in the EU Statistics on Income and Living Conditions (EU‐SILC) database.
Statistical (or synthetic) matching is a broad term used to describe the fusing of two datasets. In this context, the datasets are of households sampled from the same population. The usual approach is to define one data set as the recipient (in this case, EU‐SILC), and the other as the donor (HBS) (see Figure 1). The recipient data set contains a variable Y (e.g., on material deprivation), which is not found in the donor data set, while variable Z, expenditure, is only contained within the donor data set. The aim is to use information contained within the set of variables common to both datasets, X, for example, age, sex and income, to link records from the donor to the recipient. Therefore, expenditure is linked to EU‐SILC, which contains information on income, material deprivation and work intensity.
Figure 1Matche
FurtherTonkin
The deexaminprominthree mwere inhigher: being lo(11%) aFigure 2
Figure 2Poverty
1 ed dataset w
r details of (2013).
gree of oved, with tent. In the measures, wn poverty indespite theower (24%)and almost2)
2 y measures
with EU‐SIL
the statist
erlap betwhe differenUnited King
while 12% wn all three. e proportio), a similar double th
overlap, G
Chapter
LC and HBS
tical match
ween the thnce betweegdom, 35% were in poveIn Germanyon of peoplproportion
he proportio
ermany an
r3Monetary
48
data
ing techniq
hree poverten the Uniof people eerty in two oy, the degree in povertn were in pon were in
d United Ki
yPoverty
ques used c
ty measureited Kingdoexperiencedor more of tee of overlaty on at leaoverty on tn poverty o
ingdom, 20
can be fou
es varies acom and Ged poverty inthe measurap betweenast one of ttwo or moon all three
10
und in Web
cross the cermany parn at least onres and just n the measuthe three mre of the me (almost 4
bber and
countries rticularly ne of the over 2% ures was measures measures 4%), (see
Recommtogethestatistictheir inwhere p
3.2.8
160. Thmeasur
3.2.8.1
161. Thcomparmeasurpopulatemploy
162. Themployas “thopotenticonsumdefinitioindividuthemse
mendation er, it is reccs consider tersection, possible.
Keymea
he final parrement issu
1 Self‐e
he measurred with emrement. Thitions, a suyment as on
he first chayed. The Intose jobs whal for pro
mption is coons used uals’ self‐claelves as self‐
8: Given ommendedexamining taking adv
asuremen
rt of this sees affecting
employme
rement of mployee ins is a particubstantial nne of their m
allenge is pternational here the reofits) derivonsidered for nationassification.‐employed,
Chapter
the advand that, whepoverty mevantage of r
ntissues
ection on wg income an
entincom
self‐emploncome, in tcular issue wnumber of main source
perhaps in Labour Orgemuneratioved from tto be partal employm. For examp but are tre
r3Monetary
49
ntages of cere data aveasures basregister dat
welfare meand consump
me
oyment incoterms of bowhen incomthose in
es.
the correcganisation (on is directthe goods t of profitsment or taple, directoreated as em
yPoverty
consideringvailability aled on incomta and/or st
asures briefption expen
ome repreoth definitime is used topoverty m
ct identificaILO, 1993) tly dependeand serv
s)”. Howeveax law, wrs of limitedmployees for
g multiple llows it, come and exptatistical m
fly examinediture.
esents a paions and tho measure may have i
ation of thdefines selent upon tvices produer, this mahich may d‐liability cor tax purpos
welfare mmpilers of penditure asatching tec
es some of
articular chhe practicapoverty as income fro
hose who af‐employmethe profits uced (wheay differ frdiffer agaompanies mses.
measures poverty s well as chniques
the key
hallenge alities of in some om self‐
are self‐ent jobs (or the re own rom the in from
may view
Chapter3MonetaryPoverty
50
163. The next challenge comes in defining and identifying self‐employment income. While the self‐employed will usually pay themselves a wage, this cannot usefully be compared with the earnings of employees. This is because self‐employed workers frequently take a small salary in favour of taking dividends or putting profits into the company.
164. Clear definitions in survey questions are vital, as what individuals may think of as their income may differ from what is considered income under the definitions used (e.g., those from the UNECE Canberra Handbook, 2011). Additionally, self‐employed individuals may not be able to provide an accurate estimate of their income, particularly if they have not yet filed (or are not required to file) their end‐of‐year accounts with the tax authorities.
165. Where possible, asking about profits and other data typically required for tax purposes is the recommended approach for survey data collection. Where a respondent has not prepared accounts for the tax authorities, the alternative approach is to collect data on any earnings from their business, plus any money that has been drawn from their business accounts for personal use.
3.2.8.2 Goods/servicesproducedforownconsumption
166. Home production for own consumption refers to the goods or services that are produced within the household for the household’s own consumption, rather than for sale or exchange. The estimated market value of the goods/services is included as part of households’ self‐employment income, less any expenses incurred in their production. The value of these goods and/or services should also be counted as part of households’ consumption expenditure. In practice, most services (with the exception of housing services from owner‐occupied dwellings) are not included in operational definitions of income and consumption expenditure, whereas goods should be.
167. The relative contribution of goods produced for own consumption can vary considerably between economies. In some countries, the value of such goods and the proportion of households producing them may be negligible. Where that is the case, they are often excluded from income statistics. However, in other economies, particularly agricultural, the value of own consumption may be substantial for many households. In such cases, it is important to capture this information within welfare in order to ensure the accurate measurement of households’ poverty status.
168. Two particular examples of home production for own consumption are described in further detail below: housing services from owner‐occupied dwellings, and consumer durables.
3.2.8.3 Housingservicesfromowner‐occupieddwellings
169. Owning your own house or apartment in effect provides you with housing services, which should be considered as part of both income and consumption. The value of such services is estimated as being the market rent for a similar property, less the costs incurred by the household in their role as landlord. These housing services should feature in both income (increasing the level of household resources) and consumption expenditure (contributing to the household’s economic well‐being). Including net imputed rent is
Chapter3MonetaryPoverty
51
particularly important when making comparisons of poverty across countries, where rates of home ownership can vary considerably.
170. Where there is an established rental market, rental equivalence is generally considered the preferred approach for estimating imputed rental values. The basic econometric method that is used is hedonic regression with the attributes of the dwelling used as covariates. If there is selection bias, a Heckman correction may be applied, with a model for the housing tenure and a model for the imputation of the values. An alternative approach is to use cell‐based mean imputation, which is typically referred to as the stratification method.
171. Where rental markets are less well‐developed, one commonly used alternative to the rental equivalence approach is the user cost method, based on the estimation of the costs incurred for homeownership by foregoing the opportunity to invest in financial assets from which real income flows are created in the form of income from interest and dividends. However, research by Garner and Verbrugge (2009) shows a considerable divergence between user costs and net rents in the US. A further alternative is self‐assessment—effectively asking how much you would have to pay if, instead of owning your home, you had to rent it. These methods are summarised in the table below.
Rental markets
Developed Under‐developed
i) hedonic regression (with/without Heckman correction)
iii) user cost method
ii) stratification (cell based imputation) iv) subjective self‐assessment
172. Box 3.6 summarises some research carried out looking at the measurement of imputed rent using EU‐SILC data (Törmälehto and Sauli, 2013), while Box 3.9 describes practices across OECD members.
Chapter3MonetaryPoverty
52
Box3.6 ImputedrentinEU‐SILC:2007‐2010
In the EU‐SILC guidelines, imputed rent is included as a variable, but does not form part of the main measure of disposable income used to calculate at‐risk‐of‐poverty rates—primarily due to concerns regarding data quality and comparability. Each country estimates gross imputed rents in its own preferred way. Based on the 2010 data, the most common methods are stratification and regression. Five countries used the Heckman correction while the user cost method was applied in three to four countries. Table 1 below shows the share of market renters in each country, along with the imputation method used.
Table 1 The share of market renters (percentage from population)
2007 2008 2009 2010 Imputation method
RO 1.0 0.9 0.8 1.1 Stratification
MT 1.4 1.4 Stratification
LT 1.2 1.4 2.1 1.1 Stratification
BG 2.2 1.6 2.1 2.2 Stratification
PL 2.6 2.2 2.1 2.4 Regression
HU 2.7 2.6 2.2 2.4 Regression/subjective self‐assessment
EE 4.4 2.9 2.5 2.6 User cost
SI 5.5 4.9 4.1 5.0 Stratification
CZ 4.8 5.0 5.4 5.0 User cost, subjective self‐assessment
LV 5.7 6.6 6.3 6.7 Log‐linear regression
IS 5.7 6.8 7.9 10.4 User cost
ES 7.5 8.1 8.2 8.7 Stratification/subjective self‐assessment
SK 9.2 9.1 8.8 8.4 User cost
CY 9.9 10.6 10.3 Regression with Heckman correction
NO 10.3 9.8 10.4 10.9 Stratification
FI 9.8 10.1 10.4 10.1 Stratification
PT 9.6 11.3 10.9 12.8 Regression 2008‐
IE 8.7* 9.3 11.3 Stratification
UK 8.2 9.3 12.4 11.9 Regression with Heckman correction
IT 12.8 13.1 13.3 14.0 Regression with Heckman correction
EL 17.9 17.9 18.0 18.2 Stratification/subjective self‐assessment
BE 18.6 18.4 18.5 19.6 Regression with Heckman correction
FR 20.3 19.3 19.8 20.2 Regression
LU 19.7 19.4 22.3 27.6 Regression with Heckman correction
AT 28.7 27.5 27.7 26.7 Regression
SE 28.3 30.2 29.8 28.7 User cost
NL 33.1 32.2 31.1 32.5 Regression
DK 32.9 33.5 33.7 33.2 Stratification
DE 38.2 39.0 38.9 39.7 Stratification
* Self‐assessment in 2007 Source: Törmälehto and Sauli (2013). Note: Countries arranged by the share of market renters in 2009.
Törmälehto and Sauli (2013) showed that, overall, the inclusion of imputed rent in reported income led to a reduction in the at‐risk‐of‐poverty rate in the majority of countries. When looking at poverty by age, including imputed rent has the effect of lifting older people from
Chapter3MonetaryPoverty
53
poverty (Figure 1).
Figure 1 Income poor by age, 2009
However, the authors also establish that the data quality, completeness and transparency of the estimation methods in the EU‐SILC have shortcomings. Consequently, they conclude that further methodological studies and improvements in data quality are necessary; meaning that disposable income including imputed rents cannot yet replace the current concept of cash disposable income as the primary income measure.
Recommendation 9: Due to the challenges associated with measuring housing services from owner‐occupied dwellings and the variation in methods employed across countries, it is recommended that such services be excluded from the main poverty indicators used for international comparison. However, for national purposes, compilers of poverty statistics may find it useful to consider supplementary measures including imputed rent, or take account of home ownership in other ways, such as using an after housing costs measure.
To better aid international comparison in future, as well as the targeting of resources at the national and international level, it is recommended that international organisations develop new guidelines on the measurement of imputed rent for inclusion in poverty and inequality statistics.
Source: authors’ elaborations from the EU-SILC users’ databases 2007–2010 (March 2012).
Including net imputed rents Excluding net imputed rents
Chapter3MonetaryPoverty
54
3.2.8.4 Consumerdurables
173. Household consumer durables services refer to the imputed value of services provided by household‐owned cars, washing machines, refrigerators, clothes, etc.
174. Such items are typically bought at a point in time, and then consumed over a period of several years. In theory, consumption should only include the amount of a durable good that is consumed during the year. This can be measured by the change in the value of the asset during the year, plus the cost of locking up one’s money in the asset.
Recommendation 10: In practice, because of the challenges involved in measuring the value of services by household consumer durables, they are excluded from the operational definition of income set out in the Canberra Handbook (2011). For the same reason, they are also excluded from the measurement of consumption expenditure in practice. It is therefore recommended that the same practice apply for the purpose of internationally comparable poverty statistics.
3.2.8.5 Transfersbetweenhouseholds
175. Transfers between households can have a significant impact on the economic well‐being of the households that receive them, as well as those which pay them. Such transfers may include financial support to students or young adults living away from the home, as well as payments from family members working abroad to the rest of their family in their home country (remittances). Family support payments (such as alimony and child and parental support) also fall into this category.
176. The OECD framework for income, consumption and wealth statistics (OECD, 2013a) highlights that inter‐household transfers are:
Given without an expectation of repayment (similar to any current transfer).
Given with the aim of supporting current consumption. This is related to the classification of a specific economic flow between households as income received (when money, goods or services are used immediately or in the short‐term) or as an increment of wealth (when saved or comprising a capital item such as a
consumer durable).
Often made regularly (i.e., anticipated or relied upon by the recipient household). Regular inter‐household transfers include regular alimonies, child and parental support payments, either voluntary or compulsory. Inter‐household transfers can be donated either by family members or by other persons not living in the recipient household.
177. While regular inter‐household transfers are included as income, any regular transfers in kind (such as food) are additionally counted as part of income and consumption expenditure by the recipient household. If the transfer between households takes the form of consumer durables or assets, are intended to support the purchase of an asset, or if it is expected that the majority of the value will be used for saving (or paying off debt), then the transfer is not considered income, but rather a capital transfer.
Chapter3MonetaryPoverty
55
178. Box 3.7 provides an example of the impact of remittances on poverty in Eastern Europe and Central Asia.
Box3.7 RemittancesandpovertyinEasternEuropeandCentralAsia
Eastern Europe and Central Asia show high rates of labour mobility, which is apparent in a number of waves that have occurred since the early 1990s. The first waves of migration were often associated with movements of people to newly independent (or other) states of national origin, as well as waves of refugees and internally displaced people fleeing violent conflicts (especially in the Western Balkans). Subsequent waves of migrants were driven by economic considerations—especially the lack of productive employment opportunities at home and better possibilities abroad. As a result, labour migration and associated remittances flows have become extremely large in the region—three out of the top five remittance‐receiving countries in 2014 (by the remittance inflow to GDP ratio) were Tajikistan (41%), Kyrgyzstan (30%) and the Republic of Moldova (25%). Armenia, Georgia, and Bosnia and Herzegovina report remittances‐to‐GDP ratios in excess of 10%. For a number of these countries, remittances come close to export receipts in size, and dwarf FDI and ODA flows. They deeply affect economies and individual families alike.
A number of countries track the poverty reduction impact of remittances via data collected through household budget surveys. Kyrgyzstan’s National Statistical Committee regularly reports poverty rates with and without remittances, including a breakdown by regions. These data show a significant impact of remittances, reducing the national poverty rate from 36% to 30%. This impact is even larger for extreme poverty, which drops from nearly 8% to 1%) and for less developed southern regions, like Jalal‐Abad, Batken, and Osh.
In the Republic of Moldova, the National Bureau of Statistics (NBS) supplies data to the Ministry of Economy and Trade, which is responsible for poverty monitoring. NBS publishes only structure of incomes, which includes remittances. The Ministry of Economy publishes poverty data with and without remittances in its annual poverty monitoring reports. In 2014, remittances reduced national poverty rates from 26.7% to 11.4%. These data also show that remittances reduce poverty in villages from 35.3% to 16.4%.
In Armenia, the National Statistical Service publishes the share of remittances in household income (which on average was as high as 10.4% in 2014), but do not report poverty rates without remittances. The Ukraine State Statistical Service reports the share of a broad category “money transfers from relatives and other persons; other cash incomes” (which was 7% in 2014).
Such data face some methodological questions. Respondents could be reluctant to report remittances received from abroad (although the large scale of migration and remittance flows could reduce such concerns). Remittances be received irregularly and not always within the reporting timeframe of the survey. An IMF study using household level data for Armenia found systematic under‐reporting of remittances by about 30%.
The poverty reduction effect of remittances should be treated with caution, as simple “before and after remittances” poverty calculations do not necessarily reflect other dimensions of migration. Among other things, they do not take into account opportunity costs—back home migrants would find some jobs, lower paid than abroad, but nevertheless earn some money. Remittances may also reduce labour supply of households, consequently
Chapter3MonetaryPoverty
56
reducing incomes of families.
Source: “Labour Migration, Remittances and Human Development in Central Asia” Central Asia Human Development Series (UNDP, 2015)
3.2.8.6 Socialtransfersinkind
179. The Canberra Group Handbook (2011) defines social transfers in kind (STIK) as goods and services provided by government and non‐profit institutions that benefit individuals but are provided for free or at subsidised prices. The Handbook recommends that where possible, the value of social transfers in kind should be added to household disposable income to create a measure of adjusted disposable income. Similarly, social transfers in kind added to consumption expenditure provide a measure of actual final consumption—the
total value of all goods and services used by the household.
180. Taking social transfers in kind into account is particularly important when make cross‐country poverty comparisons. This is because in country A, certain services may be largely provided by the state, free at the point of use, whereas in country B it may be necessary to pay for those services directly. This means that (all else equal) someone with the same disposable income (or consumption expenditure) in country A would have a higher standard of living than in country B. Social transfers in kind are also very important for measuring economic well‐being within countries. Since one of the major policy aims in this area is often to make access to important services more equal, the distributional impact of social transfers in kind is generally progressive. Additionally, over time, changes in the level and form of these transfers, as well as among recipient groups, can distort cross‐county comparisons if STIK is omitted.
181. Despite this, social transfers in kind are commonly excluded from measures of income and consumption expenditure, due to the challenges associated with measurement, though a number of countries have produced some estimates, at least on an experimental basis. In addition, one potential drawback of including social transfers in kind within a welfare measure is that they may mask underlying disadvantage, though this depends to a large extent on the valuation method used.
182. Box 3.8 provides an example of income poverty measures from the United Kingdom and Finland, two countries that do produce distributional analyses of the impact of social transfers in kind (Tonkin et al., 2014). Another example is the inclusion of in‐kind benefits since 2011 within resources for the Supplemental Poverty Measure published by the US Census Bureau (see Renwick & Fox, 2016). Box 3.9 describes practice across OECD member countries.
Recommendation 11: As accounting for the value of social transfers in kind is not yet common practice, it is recommended that they be excluded from indicators used for international poverty comparisons (at least for now). It is also recommended that statistical compilers consider developing methods for including these transfers in income and consumption expenditure statistics, and invest in learning from international best practices, so that future international comparisons may be based on data in which the effects of these transfers are included. To assist with this, guidance for national statistical offices should be developed by international organisations.
Chapter3MonetaryPoverty
57
Box3.8 PovertymeasuresincludingsocialtransfersinkindinUnitedKingdomandFinland
Both the United Kingdom and Finnish national statistics offices regularly publish statistics on the distribution of income including the value of social transfers in kind (STIK), through imputing the value of in kind benefits to each country’s household budget survey data (e.g., ONS, 2016).
Both countries estimate the value of education services received by households using an “actual consumption” approach, in which an attempt is made to assess the value of education services for the households that directly benefit from these services. This approach relies on information produced by various government departments and agencies on the cost per full‐time equivalent pupil or student in different stages of education.
In the United Kingdom, the “actual consumption” approach used for assessing the value of education services is considered to be less appropriate for health care as it implies that people who are ill are better off than healthy people with the same disposable income. An “insurance value” style approach is therefore applied, under which the benefit is assigned to all households. Overall, it accounts for 60% of the total value of social transfers in kind in the United Kingdom’s 2012 household budget survey data. In Finland, the value of healthcare services is assessed using an actual consumption approach in their regular statistics, although—for the purpose of the analysis presented below (Tonkin et al., 2014), an “insurance value” approach was used to provide comparability with the United Kingdom.
In both countries, education and healthcare make up the vast majority of social transfers in kind reported, though some other forms are also estimated, notably for travel and housing subsidies in the United Kingdom and for social care in Finland (Figure 1).
The distributional impact of social transfers in kind in the United Kingdom is broadly progressive, with the bottom and second quintiles receiving the equivalent of EUR 9,200 per year, compared with EUR 6,600 received by the richest fifth. This pattern reflects the demographic profiles of the different quintiles. By contrast, in Finland, the middle and second quintiles have the highest average values; together they account for almost half (46%) of the total value of these transfers. The poorest fifth of households received the equivalent of EUR 6,100 from social transfers in‐kind, while the average in the fifth quintile group was close to EUR 6,900.
Figure 1 Social transfers in kind by household equivalent income quintile group, Finland and the United Kingdom, 2012
Chapter3MonetaryPoverty
58
Figure 2 compares poverty rates based on adjusted disposable income, including social transfers in kind. The final column uses the so‐called simplified needs adjusted (SNA) equivalisation scale (Aaberge et al., 2013), which was specifically designed for using with social transfers in kind. The OECD‐modified scale is designed for equivalising cash income and is not necessarily appropriate when STIK are included in the income measure. For example, the OECD‐modified scale assigns a smaller value for children than for additional adults, based on assumed needs. However, young children have a high need for education services and also comparably higher needs for healthcare (though less than for older people). Therefore, applying a standard equivalisation scale risks overstating the living standards of households with young children.
Figure 2 Relative at‐risk‐of‐poverty rate, United Kingdom and Finland, 2012
0
2
4
6
8
10
12
14
16
18
Disposable income (OECD-modified)
Adjusted disposable income (OECD-
modified)
Adjusted disposable income (SNA)
UK
Finland
Percentages
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
I (Bottom) II III IV V (Top) All
Average
per household (euro per year)
UK: Other, includingtravel and housingsubsidiesUK: Education
UK: Healthcare
Finland: Social Care
Finland: Education
Finland: Healthcare(IC)
Finland: Healthcare(AC)
Chapter3MonetaryPoverty
59
The impact of including social transfers in kind in this indicator is quite noticeable, particularly for the United Kingdom. Including these transfers while retaining the OECD‐modified equivalisation scale reduced the headline at‐risk‐of‐poverty rate to 9.7% in the United Kingdom (a 39% decline). This was around 1 percentage point lower than the Finnish rate for the same measure (10.8%; a relative reduction of 17.6%).
Box3.9 ThemeasurementofimputedrentsandsocialtransfersinkindintheOECD
The inclusion of imputed rents and social transfers in kind in national measures of income is of great importance to poverty and income distribution studies, and for guiding policymaking.
In 2001, the first edition of the Canberra Group Handbook included imputed rent in the conceptual definition of income but not in the operational one, mainly due to methodological concerns and the lack of harmonised and comparable data. However, in 2011, in the 2nd edition of the Handbook, the operational definition of income was broadened to include the net value of owner‐occupied housing services in the recommended income definition to be used for international comparisons.
Both editions of the Handbook acknowledged the importance of adding the value of social transfers in kind to household disposable income, to create a measure of adjusted disposable income. The 2011 edition of the Handbook also stressed that “the development of comparable estimates of STIKs should have high priority if the accuracy as well as the international comparability of income distribution statistics is to be improved”.
In practice, however, due to measurement challenges and methodological concerns, available international evidence on levels and trends in income inequalities and poverty usually keep relying on the concept of household disposable cash income, thus ignoring the services from owner‐occupied dwellings and the services that governments provide to households. At the European level, imputed rent and social transfers in kind are not included in the standard definition of income underpinning the main indicator of risk of poverty. Similarly, they are not included in the income definition underpinning the indicators of the OECD income distribution database.
In 2015, the OECD sent a questionnaire to its network of income data providers, to collect information on what statistical offices and other data producers have done, are doing, and are planning to do in terms of including imputed rents and social transfers in kind in their measures of income inequality. To date the OECD has received replies from 27 countries, which show the interest that national statistical offices give to this area of work. A brief overview of the main results of the OECD questionnaire is provided here, while a more detailed analysis can be found in Balestra and Sustova (2017).
Only 3 out of 27 OECD countries (Canada, Republic of Korea, and United States) do not compute imputed rent. Most countries produce and disseminate such estimates annually,
Chapter3MonetaryPoverty
60
with 2013 and 2014 being reported as the most frequent “latest year of estimate”. The rental equivalence approach is used by the large majority of countries, and implemented through different methods—including the stratification method, the hedonic and linear regression methods, and the subjective method, as well as a mix of the above. Most countries do not plan to change measurement approaches/methods in the near future, either because the current one is delivering satisfactory results or because they want to avoid breaks in the relevant data series.
Half of the countries producing estimates of imputed rent include them in national definitions of income; five include imputed rent in the main national concept of income; and seven include it in secondary or alternative ones. The inclusion of imputed rent in the operational income definition produces different effects on poverty rates: poverty decreases in a small majority of countries, while it increases in Austria, Finland, France, Mexico, and Norway. The impact also varies across different population groups.
Only ten of the countries that answered the questionnaire are computing estimates of social transfers in kind (Australia, Austria, Denmark, Finland, France, Japan, Mexico, the Netherlands, Norway, and Sweden). The majority of countries produce and disseminate such estimates regularly—two to five years is the most frequent periodicity for publication of the estimates. All countries include healthcare benefits in their estimates of STIKs; almost all include education and childcare, and a majority include long‐term services for the elderly. Two countries (Australia and France) include social housing in the estimates, while Norway includes subsidies for public transport, social services targeted towards disadvantaged individuals and culture. As for the valuation methods used, countries can be grouped in three groups: those that use the average cost of production approach, those that use the average cost of provision approach, and those that use a mix of these two approaches. Denmark, Finland and Mexico allocate the value of these transfers to beneficiaries (actual consumption approach); Japan, Netherlands and Norway allocate the value of services equally among those having certain characteristics (insurance value approach); while the remaining four countries (Australia, Austria, France and Sweden) use a combination of approaches. In most countries, these transfers are attributed to the individual beneficiaries, although in a few countries they are attributed to the household as a whole. Only 4 out of the 10 countries that produce estimates of STIKs include their value when computing household income. Estimates of the size of these transfers vary from 7.1% of household cash disposable income in Mexico to up to 44% (if only the part of public consumption that may be individualised is distributed) or 62% (if total public consumption is distributed to individuals) in Denmark. The average share is however lower—around one fifth to one fourth of household cash disposable income. Half of countries reported a decrease in national poverty rates due to the inclusion of STIKs in the definition of income.
The 2016 EU‐SILC ad hoc module on access to services could potentially help to make the imputation of STIKs more accurate. The module focuses on the affordability of services, and on the unmet needs for such services; questions on the cost of these services paid for by households are also included, which could help identify the exact amount of social transfers in kind to allocate to a particular household. The ad hoc module considers the following services: childcare, formal education and training, lifelong learning, healthcare and professional homecare.
3.2.9
183. Coaccumucrashesthan thallows ichange
184. Cohighly cwealth less thaper cen
FigureNetwe(Mean
S
185. Wfor incoexamplthose wthose wworse‐ointo mohousing
186. Awith hig
Wealth
ompared wulated savins in investmheir incomeindividuals s to income
ompared wconcentrateby wealth dan 1000 eurnt of househ
3.1 ealthofprianddecile
Source: Hou
While wealthome and ce, considerwho have swho lack thoff than theoney, its exg improvem
dditionally,gh enough
with incomengs and invment or houe, or they mto smooth e.
with the dised towards decile in Auros and theholds had ne
ivatehouspoints)
usehold Fin
h cannot beonsumer ering wealthufficient whis buffer. Be former. Wxistence maments, moto
it allows tincome or
Chapter
, wealth (a vestments, using markmay consumconsumptio
stribution othe top of
ustria in 201e median wet assets in
seholdsin
ance and C
e used on itxpenditureh makes it ealth to suBoth groupsWhile some way allow peor vehicle pu
the identificconsumpti
r3Monetary
61
stock measalthough iets. Houseme less thaon over tim
of income f the distrib10. The pooas around 8excess of 5
Austria,2
onsumption
s own as a , it can propossible topport theirs experiencwealth is heeople to bourchases, an
cation of thion levels t
yPoverty
sure), is moit can dropholds can uan their incme and to p
and consumbution. As arest 10% of86,000 euro518,000 eur
014
n Survey (H
measure ofovide impoo distinguisr living stance low incoeld in assetsorrow to finnd so on).
he “asset‐poo achieve a
ore stable op dramaticause wealth come, and rotect them
mption, wean examplef householdos. By contros.
HFCS)‐Austri
f poverty, uortant additsh (among ndards for ame, but ths that are nnance expe
oor only”, ta given min
ver time, really in the to consumthus save.
m from une
ealth is evee, Figure 3.1ds had net arast, the ric
ia (2014).
used alongstional insigthe incom
a given pere latter areot easily coenditures, (
that is, hounimum stan
eflecting case of
me more Wealth xpected
en more 1 shows assets of chest 10
ide data hts. For
me‐poor) riod and e clearly onverted e.g., for
useholds ndard of
Chapter3MonetaryPoverty
62
living at this moment, but do not have enough assets to protect them should their resources for consumption fall suddenly.
187. Collection of microdata on the distribution of household wealth is not as well established in statistical offices as is the collection of microdata on the distribution of income or consumption expenditure. However, an OECD expert group has developed international guidelines for wealth microstatistics (OECD, 2013b), and increasing numbers of countries now collect such data, sometimes through their central bank.
188. To be used in the measurement of poverty, such wealth data need to be compatible with poverty data from other sources. In practice, this is challenging, particularly in countries where data collection relies on household surveys rather than registers, due to the substantial burden placed on survey respondents. In such cases, it may be possible to utilise statistical matching techniques to create synthetic datasets containing all the variables of interest (see e.g., Tonkin, Serafino and Davies, 2016). An alternative approach which has been used is to proxy liquid financial wealth using other survey variables, such as the EU‐SILC variable on capacity to face unexpected expenses (see e.g. Morrone et al, 2011; or Tormalehto et al., 2013).
Recommendation 12: While wealth is an important factor to consider alongside income or consumption in assessing poverty, it cannot be used as a measure of poverty on its own. It is recommended that countries invest in developing wealth statistics that can be assessed alongside other welfare measures, with the long‐term aim of being able to consider jointly the distribution of income, consumption, and wealth, in order to provide a complete picture of individuals’ economic well‐being. This should be possible when registers and other administrative data sources are available to producers of statistics. Alternatively, statistical matching techniques should be utilised where income (or consumption) and wealth are not available through the same survey source.
3.3 S
3.3.1
189. Apovertyline”). conside
190. Th
191. Th
192. Inshould
193. Arelative
3.3.2
194. Stpovertylines arspace. transpapurely tmost comonetaaccepta
195. Thabsolutcomposcountriclothing
Settinga
Absolute
s describedy) if its resoThe povertered necess
here are thr
The absominimum
The relat
The subje
he first two
n setting a take the fol
Is there a
Is the dequality st
Has the dregion? (then they
Does the
lmost all coe poverty lin
Absolute
trictly speay that does re by far thThey are arent compato changes ommon appary value ofable standa
here are sote poverty sition of thies and regig and hous
apovert
eandrela
d above, anources are ty line repary to satisf
ree basic ap
olute poverm”, Hagenaa
tive poverty
ective pove
o of these ap
poverty linllowing poin
a clear defin
efinition bastandards?
definition b(If calculatioy may still q
definition
ountries witnes and pro
epoverty
aking, absonot depende most comuniversally arisons ovein the distproach to ef a set of grd of living.
ome imporlines usingis basket ofions. The resing in the
Chapter
tyline
ativepov
n individualess than t
presents thfy the hous
pproaches t
rty line (or ars and De V
y line (or “ha
rty line (or
pproaches a
ne, the defints into acc
nition of the
sed on an e
been applieons have bqualify);
produce inf
th official mduce pover
ylines
lute povertd on the avemmonly useused in lo
er time whetribution ratestablishingoods and s
tant practig such an f “essentialesources thearly part
r3Monetary
63
vertyline
al or housethe value oe aggregatehold’s bas
to establish
“having leVos, 1988)
aving less th
“feeling you
are the prim
inition of wcount:
e relevant st
existing sou
ed more thabeen carried
formation t
measures of rty indicator
ty thresholerage conded approacow and mere changesther than tg an absolutervices con
cal challengapproach. ” goods anhat might bof the 20t
yPoverty
es
ehold is claof a given mte value ofsic needs.
ing a pover
ss than an
han others”
u do not ha
mary focus o
what can be
tandard and
urce of info
an once, idd out only
hat is usefu
poverty bars according
ds may beition of theh for identiiddle‐incoms in measuro a movingte poverty nsidered ess
ges associaFirst, it is
nd services be considereth century
ssified as pmonetary tf all the g
ty line:
objectively
”)
ave enough
of this chap
e considere
d its units o
rmation tha
eally for thonce but fo
ul for public
se these ong to these li
e based one populationifying the pme countriered povertyg poverty cline involvesential to a
ated with ts unrealistias fixed oved essentiaare likely t
poor (or athreshold (“oods and
y defined a
to get by”)
pter.
ed “good p
of measurem
at meets m
he same couor many co
policies?
n either absines.
n any definn. Absolute poor over ties. They aly can be attut‐off. Perhes determinchieve a m
he construc to mainter time andal in terms to be subs
t risk of “poverty services
absolute
practice”
ment?
minimum
untry or ountries,
solute or
nition of poverty ime and llow for tributed haps the ning the
minimally
ction of tain the d across of food, tantially
Chapter3MonetaryPoverty
64
different to what is deemed vital today (as a result of technological advances and general improvements in living conditions) in order to reach a decent standard of living.
196. This raises the important question of how frequently an absolute poverty line should be updated. Here the trade‐offs are clear: the threshold should be fixed long enough to be able to discern underlying changes in poverty; and it must be updated often enough so that the standard is reasonably consistent with prevailing circumstances. Absolute poverty lines are often held constant over a long period, and then updated to reflect changing living standards. After updating, comparisons are typically not made across the two standards. Instead, each distribution is evaluated at the new, updated poverty line.
197. The United States poverty line has remained fixed (in real terms) since 1965 (see Box 3.10); the nominal poverty line is adjusted for inflation. The World Bank’s main poverty standard was updated in 2015, and all income distributions back to 1981 were re‐evaluated at the new line.
198. Although absolute poverty lines are most commonly used in the developing world, there are a number of reasons for also considering such an approach in high‐income countries. First, although the proportion of the population in absolute poverty in such countries may be very small, it is important to understand the characteristics of those who are living in poverty, in order to effectively target welfare policies25. Additionally, an absolute poverty line is effective for use in evaluating poverty within a country over short‐to‐moderate spans of time, or across two countries when they have roughly similar levels of development. This approach may be harder to justify over longer periods of time or in a comparison of countries with very different levels of development.
3.3.2.1 Settingabsolutepovertylines:thecostofbasicneedsapproach
199. The ways in which absolute poverty lines are set vary considerably across countries. As stated above, the cost of basic needs approach is most commonly used, particularly in developing countries, but variations in the way the approach can be applied multiply with each step. The basic method involves estimating the cost of acquiring enough food for adequate nutrition, using a pre‐defined number of calories per equivalised26 adult per day and then adding the cost of other essentials such as clothing and shelter.
200. The process of defining a poverty line using the cost of basic needs approach is set out in more detail below.
3.3.2.1.1 Specify a food poverty threshold
201. Food poverty lines are based on minimum nutritional requirements. People are counted as food poor if the nutritional content of the food(s) they consume is below a prescribed threshold. As a simplifying assumption, most countries use dietary energy
25 The use of absolute poverty lines in developed countries make the challenges of adequate population coverage of statistics, discussed in section 3.1.5, particularly acute. Many of those who are in absolute poverty may not be covered by traditional household surveys. 26 Equivalisation is the process of adjusting a welfare measure to account for household size and composition, taking into account any possible economies of scale. The use of equivalence scales is discussed in Section 3.3.4.1
Chapter3MonetaryPoverty
65
(calories) as a proxy for overall nutritional status—i.e., if people get enough calories in their diet, it is assumed that they obtain adequate protein and other essential nutrients.
3.3.2.1.2 Food basket construction and food poverty lines
202. To determine the cost of the food basket, there are two different basic methods: constructing an explicit food basket and then pricing it, or estimating the cost of the food basket without listing its contents.
A. Constructing an explicit food basket. Under this approach, an explicit bundle of food by item and weight (for example, meat—0.25 kg, sugar—0.03 kg, etc.) is seen as needed to provides a total close to the specified threshold in kilocalories per equivalised27 adult per day. The conversion is made through a so‐called food composition table, which varies across individual countries to reflect their individual situations. The composition of the food basket depends on the choice of reference population. Since the aim is to identify and count the poor, the reference population is usually some lower percentile of households according to their equivalised adult consumption expenditure distribution. The choice of the percentile cut‐off point is usually guided by the most recent poverty incidence estimates, which infers that the reference population should be similar to the poor population.
B. Price per kilocalorie. This method avoids constructing a food basket by instead calculating the total expenditure and total caloric content of all the food consumed by the reference population. The ratio between the two totals is a price per kilocalorie estimate. When this figure is multiplied by the energy threshold, it provides an estimate of the food poverty line. Once a price‐per‐ kilocalorie estimate is calculated, food poverty lines for different calorie thresholds are easily calculated. However, this approach requires the more extensive conversion into energy equivalents, as there are more food commodities consumed by the reference population.
3.3.2.1.3 Calculation of the total poverty line
203. This process involves two steps: defining essential non‐food basic needs, and adding their costs to the food poverty line. Essentially, the food poverty line needs to be adjusted upward by an amount equal or proportionate to the costs of acquiring essential non‐food basic needs of a person experiencing poverty or near‐poverty. Therefore, these essential non‐food basic needs require a definition that can be measured. Developing countries generally follow one of two operational definitions or procedures:
List of specified essential non‐food needs. This list is usually created by a group of users and stakeholders in association with the national statistics office or the agency charged with producing the country's official poverty statistics. The list is exhaustive, covering items like clothing and footwear, shelter, fuel and light, household goods, health services, personal care, and education. Costs per person are assigned to each item. Hence, if the non‐food poverty line denotes the sum of the costs, then total poverty line is equal to the sum of food and non‐food poverty lines. However, the outcome is very sensitive to the contents of a highly subjective list. Additions to or
27 Equivalisation is the process of adjusting a welfare measure to account for household size and composition, taking into account any possible economies of scale. The use of scales are discussed in Section 3.3.4.1
Chapter3MonetaryPoverty
66
subtractions from this list directly affects the total poverty line. This can easily give rise to questions about why one item is included while another is not.
Engel coefficient. The most commonly used approach for setting the non‐food poverty line is based on the observed Engel coefficient (the proportion of expenditure devoted to food) for a reference group of the population. This approach consists of multiplying the inverse of this coefficient by the cost of the food basket, so that the non‐food basket cost is directly obtained from the consumption habits of the reference population. This methodology is based on the original work done by Mollie Orshansky when drawing the United States poverty lines (see Box 3.10); it is sometimes referred to as the Orshansky multiplier. As an alternative, another reference group for the construction of the non‐food poverty line, such as households with a level of total expenditure close to the food poverty line, may also be selected.
3.3.2.1.4 Advantages and disadvantages of the cost of basic needs approach
204. Advantages:
Thresholds are defined directly from surveys;
Data are comparable over short‐to‐moderate spans of time.
205. Disadvantages:
The threshold is very sensitive to the initial calculation point—changing the reference point in time for calculating the poverty threshold can change reported poverty indicators, even though the underlying tendencies remain the same;
There is the related question of how frequently to update the poverty line. But here the trade‐offs are clear: the poverty line must be fixed long enough to be able to discern underlying changes in poverty levels, and it must be updated often enough so that the standard is reasonably consistent with prevailing circumstances. Similar to any absolute poverty measurement, the poverty line is often held constant over many periods, and then updated to reflect changing living standards. After updating the line, comparisons are typically not made across the two standards;
Estimates of energy (caloric) requirements for the population are generally based on internationally agreed recommendations (FAO/WHO), but the actual energy (caloric) requirements used by the countries differ, which can lead to comparability issues;
The consumer price index as currently constructed in most countries may not reflect the consumption pattern of the reference population used in determining the poverty lines.
3.3.2.2 Settingabsolutepovertylines:thesubsistenceminimum
206. Another method used for constructing absolute poverty lines is the subsistence minimum approach, which is used mainly in CIS countries. For calculating the subsistence minimum, calculation of both food and non‐food components are needed.
Chapter3MonetaryPoverty
67
3.3.2.2.1 Calculation of food component or food poverty line
207. The subsistence minimum is defined on the basis of a minimum food basket. The minimum food basket represents a table of defined quantities of food products and contains the amount of food that is physiologically required (proteins, fats, and carbohydrates) for an equivalised adult to lead a normal life and have the ability to work, which is converted to certain amount of calories. Usually, the minimum food basket is defined by the ministry of health or other such agency. Initially the cost of each component of the minimum food basket is calculated by means of average food prices. The monthly cost of a food basket product is then obtained by multiplying a product’s monthly norm by its average price. The sum of the costs for all the food basket products therefore represents the monthly cost of the minimum food basket.
3.3.2.2.2 Calculation of the total poverty line
208. This computation involves defining essential non‐food basic needs and adding their cost to the food poverty line, to arrive at the total poverty line. As with the cost of basic needs approach, the food poverty line needs to be adjusted upward by an amount equal or proportionate to the cost of acquiring the essential non‐food basic needs of a person experiencing poverty or near‐poverty. To define these, countries typically follow one of two procedures:
1. A list of specified essential non‐food needs is compiled. This list can be created in the same way as for the cost of basic needs approach.
2. A use coefficient—the share of food expenditure in total household consumption expenditure—can be derived from an ongoing household survey. This is done by multiplying the inverse of the coefficient by the cost of the food basket. Here the reference population is usually some lower percentile of households according to their equivalised adult consumption expenditure distribution. The choice of the percentile cut‐off point to calculate the subsistence minimum or total poverty line is usually guided by the most recent poverty incidence estimates. The obtained amount represents the final value of the subsistence minimum for an equivalised adult in a given month.
209. With this approach, the total poverty line is updated by calculating the cost of the subsistence minimum on a monthly basis.
210. There are a number of disadvantages to the subsistence minimum approach:
The threshold is not defined from any survey;
The basket of goods used for the consumer price index may diverge significantly from the one used to construct the poverty lines;
Changes in prices (inflation) may not correspond directly to changes in the population’s (especially to the poor population’s) welfare. Increases in the prices for certain products in the minimum food basket automatically produce increases in the threshold (subsistence minimum), even if poor households switch to cheaper food products of a similar nature;
The seasonality of food products that are given in the minimum food basket will have an impact on the food poverty line.
Chapter3MonetaryPoverty
68
211. Box 3.10, Box 3.11 and Box 3.12 provide examples of how absolute poverty measures have been implemented in developed economies, including the United States, Italy, and the Russian Federation. In addition, the European Reference Budgets Network, a pilot project established by the European Commission, was set up in 2013 to develop a proposed common methodology for comparable reference budgets for all EU countries for different family types. The intention is that such budgets could be used for poverty measurement as well as helping EU member states design effective and adequate income support measures (Goedemé et al., 2015).
Box3.10 TheofficialthresholdsmeasureintheUnitedStates
The United States uses an official measure of absolute poverty which dates from the early 1960s. Molly Orshansky, an economist at the Social Security Administration, developed a measure of poverty, which over time has become an absolute measure, based on the cost of a minimum diet multiplied by a factor of three.
Orshansky’s “generally accepted” standards of adequacy for food were taken from the “economy food plan” prepared by the United States Department of Agriculture, which was developed for “temporary or emergency use when funds are low” and made no allowance for the possibility of eating meals outside the home.
The multiplier of three was derived from an analysis of then relatively current household survey data from the Household Food Consumption Survey conducted in 1955. This analysis revealed that food expenditures accounted for about 1/3 of after tax family income for families of three or more. Thus, the other 2/3s would cover all other expenses made by families. This pattern of expenditures reflected the situation of these particular families in the U.S. during the mid‐1950’s through early 1960’s based on household survey data. For these families, poverty thresholds were set at three times the cost of the economy food plan. Different procedures were used for calculating poverty thresholds for two‐person households and persons living alone.
The measure provided a range of income levels, or thresholds, which took into consideration different needs of children, adults, and seniors and of farm and non‐farm households (to take self‐production into account). Regional differences in living costs were not taken into account. To estimate poverty rates, these thresholds were compared to pre‐tax cash income
The resulting poverty lines were re‐evaluated every year on the basis of the changes in the cost of the food plans. Only two sets of changes have been made over the years:
In 1969, these thresholds were adopted as the official poverty measure. The method for annual updating was changed to the Consumer Price Index. For families living on farms the threshold value was increased from 70% to 85% of that of the other families.
In 1981, there were several minor changes. The thresholds were simplified by removing the distinction by gender of the household head and the distinction between farm and non‐farm families. The largest family size covered by poverty lines became nine persons or more.
Except for these relatively small changes, the poverty line has not been changed since 1965.
Chapter3MonetaryPoverty
69
The United States therefore use an absolute poverty line, which is updated only with respect to price changes.
Every year, the official poverty estimates are produced by the United States Census Bureau, using pre‐tax cash income estimates from the Annual Social and Economic Supplement of the Current Population Survey to estimate the number and proportion of the poor.
The United States official poverty measure has been widely criticized over the years. In 2010 an interagency technical working group instructed the US Census Bureau, with the assistance of the Bureau of Labor Statistics, to produce a supplemental poverty measure (SPM) that addresses many of the shortcomings of the official measure. Note that the new measure does not replace the official measure but rather provides a supplemental set of estimates. The new measure uses thresholds derived from the Consumer Expenditure Survey that are updated each year based on expenditures on food, clothing, shelter and utilities and adjusted to reflect geographic and housing tenure (i.e. owners with a mortgage, owners without a mortgage and renters) differences in housing costs. The SPM uses a resource measure that takes into account many social transfers in‐kind, tax credits and obligations and subtracts necessary expenditures including work‐related transportation, childcare expenditures and medical out‐of‐pocket expenditures. For more information on the supplemental poverty measure, see Renwick and Fox (2016).
Box3.11 EstablishinganabsolutepovertylineinItaly
The Italian National Institute of Statistics produces annually a measure of absolute poverty based on a basket of goods and services considered necessary in order for a household to avoid extreme social exclusion (basic needs).
The basket is made up of three components, the first being a food and beverage component. This is based on the calories individuals need to carry out normal daily activities. This is assumed to be invariant over time and independent of individual preferences for particular food and beverages. In order to establish nutritional levels correctly, a nutritional model defined by the National Nutritional Institute, showing daily individual diets based on sex and age, is used. This model translates the Italian recommended daily allowances into combinations of average daily food quantities. The costs of the food basket are estimated on the basis of the lowest consumer prices available for each household in Italy. The choice of using the lowest price available instead of the absolute minimum price reflects the facts that the price/cost of a good or service can vary depending on market characteristics, and that not all households have the same opportunity to buy at the same price, due to both differences in supply/availability and the mobility of individual households. Currently, this component is calculated for individuals without taking into account possible economies of scale available to households of different size.
The second component of the basket is housing. This includes a value both for the accommodation itself and the facilities it should contain. For the accommodation itself, expenditure on rents is used, making use of national regulations that link specific household
Chapter3MonetaryPoverty
70
sizes to minimum dimensions for a property. Electricity and heating costs form part of the housing measure; a TV, refrigerator, washing machine, and a non‐electrical cooker are also included in the basket of basic needs. For these goods, depreciation is taken into account, calculated on the basis of average duration of ownership (estimated from insurance data).
These two components alone do not give a complete picture of individual and household needs, as health, education, transport, and clothing expenses are excluded. For these important but extremely difficult to calculate needs, a lump‐sum residual component is also calculated. As these expenditures depend on individual characteristics and less on economies of scale, and since the individual items cannot be adequately quantified (i.e., how many and what kind of clothes an individual needs), the residual component is calculated as a percentage of expenditure on food and beverages.
The total monetary value of the three components serves as a standard reference consumption expenditure value for an Italian household that guarantees adequate nourishment, a decent dwelling, and the fulfilment of other basic needs and avoids any kind of social exclusion. The main principle is that the basic needs are the same all over the nation—adjusted for local differences due to such factors as the climate in determining heating needs. The baskets’ monetary value and the poverty threshold therefore vary by geographical area and municipality size. In the Italian approach, poverty thresholds are calculated for each household, depending on number of people and their age.
To adjust the poverty threshold for price changes over time, consumer price analytical indexes (the specific index for each good and service in the basket) for the whole community are used. Under the assumption that prices trends can differ spatially, the deflation/inflation is done by geographical area.
Box3.12 EstablishinganabsolutepovertylineintheRussianFederation
The Russian Federal State Statistics Service (Rosstat) has produced indicators for income distribution since 1970 and for poverty since 1990. Currently, Russian statistics measure both absolute and relative poverty.
The formal absolute national poverty level is assessed according to the “population with incomes below the subsistence minimum” indicator. This is calculated for the Russian Federation as a whole, as well as by the constituent entities of the country.
In accordance with Federal Law of 24 October 1997 (“On the Subsistence Minimum in the Russian Federation”), the subsistence minimum is defined as the valuation of the consumer basket as well as compulsory payments and fees.
Starting from 2013, the procedure for calculating the consumer basket and the subsistence minimum was changed in accordance with the Federal law “On the consumer basket for the Russian Federation as a whole” (3 December 2012). According to this law, the consumer basket reflects the prices of a minimum set of foodstuffs (in real terms), as well as non‐food products and services, the value of which is determined in relation to the food basket (the costs of foodstuffs constitute 50% of the total). This minimum cost food basket is designed
Chapter3MonetaryPoverty
71
to represent what is necessary to maintain human health and ensure normal activities. The basket is established for the Russian Federation as a whole, with some variation for sub‐national conditions.
The valuation of the consumer basket is based on Rosstat data on consumer prices for food and consumer price indices for food, non‐food products, and services. The value of the subsistence minimum is determined on a quarterly basis.
3.3.2.3 Setting absolute poverty lines: Estimating international povertylinesforgroupsofcountries
212. The previous sections set out the two main approaches to setting absolute poverty lines for individual countries. However, international comparisons of absolute poverty require poverty lines that can be used across groups of countries. The World Bank’s international poverty line for measuring extreme poverty (currently PPP$1.90/day using 2011 exchange rates) has historically been based on a “typical” value of a sub‐sample of the lowest of a group of national poverty lines. There are a number of criticisms of this approach, including the sensitivity of the results to the method for selecting the sub‐sample, and the quality of the data for inflating the lines vis‐à‐vis the reference year for the comparison. It also only provides a measure of extreme poverty in global terms, and is therefore unsuitable for use among CES countries.
213. Jolliffe and Prydz (2016) propose a new approach for estimating international poverty lines for groups of similar countries, based on national poverty data. Their approach is based on estimating implicit national poverty lines by combining national poverty headcounts from national sources with corresponding consumption and income distributions.28
214. Using a sample of 115 national poverty lines, Jolliffe and Prydz (2016) consider two different ways to select a reference set of national poverty lines upon which to base international lines. The first is to group countries into quartiles based on their per capita household final consumption expenditure. The second is to use their World Bank official income classification, which is based on GNI per capita. Taking median values of poverty lines for both approaches yielded very similar results (within 5%). In particular, the median for low‐income countries was PPP$1.91/day, close to the established World Bank international poverty line of PPP$1.90/day. The estimated lines from this research are set out in the Table 3.2 below.
28 A further approach to be considered for international poverty comparisons is the weakly relative poverty line (Ravallion & Chen, 2009). This is discussed in Section 3.3.3.4 below.
Table3Interna
A. Countquartile
Lowest 2
25‐50 %
50‐75% (
Highest 2
S
quartile iGNI, in p
3.3.3
215. Retype ofwhich athey bethat the(Towns
216. A expendexamplcountrycut‐off Indeed,poverty
217. It used inapproacsubsisteinterneusually regular such as
218. Rethere isinterest
219. Wthan wincidenrelative
3.2 ationalpo
try groups
25% (29)
(29)
(29)
25% (28)
Source: JolliffNote: Countrin panel A ananel B. Numb
Relative
elative povf diet, partiare customelong. Their ey are in efend, 1979).
relative poiture) distre is the Euy’s median (below whic, a level of iy line of ano
is also posn the Canadch, items aence level, t connectiobe consideinterval, n "data plans
elative povs less conct in inclusio
With a relatiwith an absce of povee income po
overtyline
Median
1.86
3.34
5.62
22.20
e and Prydz (2ies are groupd categorized
ber of countrie
Poverty
erty has becipate in thary, or at leresources affect exclud.
overty line ribution, nauropean Un(disposablech one is concome thatother distrib
ssible to dedian Markeare chosen and relati
on, tickets toered for an ew items ts" for mobi
verty lines acern about n or relativ
ive line, thesolute line. rty calculatoverty tren
Chapter
sbygroup
Mean
2.11
3.65
6.17
21.45
2016). ped based ond based on Wes in each gro
Lines
een describehe activitieseast widelyare so serioded from th
is usually aamely, a cnion’s coun) income. Tonsidered pt is above thbution havin
evelop basket basket Mthat ensuve to medo cultural oabsolute linhat have ble phones.
are most ofachieving e position.
e analysis oMoreover
ted by the nds also sho
r3Monetary
72
p
B. IncoClassif
Low In
Lower
Upper
High I
per capita hWorld Bank offup in parenth
ed as peops and havey encourageously below he ordinary
an explicit fconstant frantry‐level poThe nature opoor varies he poverty ng a higher
ket‐based rMeasure of re a standdian standaor sporting ene. When tecome mor
ften used ia minimum
of a change, if the increlative lineow trends i
yPoverty
ome fications
ncome (33)
r Middle (32)
r Middle (32)
ncome (29)
household finaficial income eses.
le who “lac the living ed or approthose com
y living patt
function of action of soverty linesof a relativeproportionline in one income sta
relative povlow incomard of livinards. Items events, andhe basket ire common
n countriesm absolute
in povertycomes of ee will remain income i
)
al consumptioclassification,
ck the resouconditions oved in the manded byterns, custo
f the incomsome incoms, which are poverty linally with itsdistributionndard.
verty threshme. In a basng above tmight incl
d other items rebased, n to society
s with highe living stan
y over time every housain the saminequality.)
Median M
1.91
3.21
5.48
21.70 2
on expenditu which uses p
urces to oband the amsocieties t the averagoms, and ac
e (or consume standare set at 60ne dictates s income stn may lie be
holds, suchsket‐based he basic nlude a highms which woas it wouldy may be ad
er incomesndard and
is less transehold doume. (In this In addition
Mean
2.23
3.87
5.61
21.19
re (HFCE) per capita
btain the menities o which ge family ctivities”
umption rd. One 0% of a that the tandard. elow the
as that relative eeds or h speed ould not d be at a dded in,
s, where greater
nsparent ble, the respect, n, if the
Chapter3MonetaryPoverty
73
incomes of poor households fall but by smaller amounts than the incomes of wealthy households, relative poverty decreases.
220. A key advantage of a relative poverty line is its conceptual clarity and simplicity of use (particularly for international comparisons). However, any relative poverty threshold is essentially arbitrary. In addition, unlike absolute poverty, the relative poverty line can increase with general increases in average incomes, if the incomes of poorer households rise more slowly than national averages. In this case, the process of poverty reduction will not be as obvious as in the case of absolute poverty. Given the advantages and disadvantages of both relative and absolute poverty measures, many countries use both types of measure (as well as subjective measures in some cases), in order to gain a richer understanding of poverty in their country.
3.3.3.1 Meanvsmedian
221. The selection of the parameter to determine the value of the poverty line influences both its measurement and interpretation. The arithmetic mean, which is relatively straightforward to calculate, shows the level of income (consumption expenditure, welfare) all households would have if incomes were equally distributed. However, this parameter can be very sensitive to extreme values of the distribution. Related to this, where microdata have been top‐coded (constrained to an arbitrary maximum, perhaps to minimise disclosure risk), this will have a negative impact on the mean, but not the median.
222. For this reason, the median income is commonly used for determining the poverty line. The median income is the household income of what would be the middle individual if all individuals in the population were sorted from poorest to richest. As it represents the middle of the income distribution, the median household income provides a good indication of the standard of living of the “typical” individual in terms of income. The median is the most stable among other measures, and it is the most appropriate choice for a log‐normal distribution (which often well approximates the distribution of income or consumption expenditure) as well as for predicting separately the effects of the economic cycle and inequality within the distribution.
Recommendation 13: In setting relative poverty lines for international comparison purposes, it is recommended that the median is used as a parameter, as it provides a more stable threshold which is less affected by measurement issues towards the top of the distribution.
223. The difference between the mean and the median can be regarded as one measure of income dispersion. In most countries, average household income will be higher than the median household income. The reason for this is that the distribution of income is usually skewed towards the lower end of the distribution. The ratio between the mean and the median can be considered a crude measure of inequality: the higher the value, the greater the inequality.
224. The poverty line calculated using the mean is very often higher than that calculated by the median; consequently, the apparent proportion of the population deemed to be at risk of poverty would typically be greater. An example, showing the impact of using mean or median income in calculating the at‐risk‐of‐poverty threshold in EU‐SILC is given in Box 3.13.
Chapter3MonetaryPoverty
74
Box3.13 Useofmeanandmedianincomeintheat‐risk‐of‐povertythresholdinEU‐SILC
Looking across the countries producing EU‐SILC data, the difference between mean and median (for the relative poverty threshold set at 60%) varies from 6.5% (in Sweden) to 27.9% (Cyprus). As a consequence, the difference between the poverty rates obtained using these thresholds varies from 3.0 percentage points for Norway to 16.9 percentage points for Cyprus (see Table 1).
Table 1 Individual at risk‐of‐poverty thresholds and rates for national populations, 2014
Poverty threshold ((OECD‐modified) equivalised
disposable incomes in Euros) set at: At risk‐of‐poverty rate
(share in total population)
60% median 60% mean 60% median 60% mean
Belgium 13,023 14,058 15.5 19.9
Bulgaria 1,987 2,344 21.8 28.8
Czechia 4,573 5,160 9.7 15.0
Denmark 16,717 18,665 12.1 17.6
Germany 11,840 13,522 16.7 22.5
Estonia 4,330 5,292 21.8 31.8
Ireland 12,101 14,234 16.4 25.0
Greece 4,608 5,327 22.1 28.3
Spain 7,961 9,243 22.2 29.4
France 12,719 14,767 13.3 21.5
Croatia 3,135 3,479 19.4 24.0
Italy 9,455 10,748 19.4 25.3
Cyprus 8,640 11,051 14.4 31.3
Latvia 3,122 3,794 21.2 31.2
Lithuania 2,894 3,585 19.1 30.2
Luxembourg 20,592 23,133 16.4 22.5
Hungary 2,707 3,074 15.0 21.1
Malta 7,672 8,574 15.9 22.7
Netherlands 12,535 13,914 11.6 16.3
Austria 13,926 15,648 14.1 19.5
Poland 3,202 3,698 17.0 24.3
Portugal 4,937 5,914 19.5 28.4
Romania 1,293 1,466 25.1 30.0
Slovenia 7,146 7,706 14.5 17.7
Slovakia 4,086 4,491 12.6 16.4
Finland 14,221 15,678 12.8 18.1
Sweden 16,272 17,331 15.1 18.2
United Kingdom 12,317 14,482 16.8 26.0
Iceland 13,492 14,712 7.9 12.0
Norway 26,265 28,181 10.9 13.9
Source: Data from EU‐SILC surveys ilc_li01 and ilc_li02. Available from http://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=ilc_li01 and http://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=ilc_li02.
Chapter3MonetaryPoverty
75
3.3.3.2 Levelofthreshold
225. Together with the selection of the threshold parameter, the choice of the fraction of the parameter can also strongly influence reported poverty levels. Not only because the higher the fraction, the greater the poverty rate; but also because the reported rate depends on the distribution/concentration of the equivalised distribution around the chosen value.
226. The most commonly used thresholds tend to be either 60% (used by Eurostat and many EU countries) or 50% (used by the United Nations and the OECD, as well as individual countries) of median equivalised disposable income. These thresholds are commonly used in part for reasons of comparability with other approaches, and in part due to pragmatic considerations such as measurement error and accepted practice.
227. The use of a single poverty line can be complemented by use of additional thresholds (obtained as a percentage of the standard threshold) in order to highlight the sensitivity of the results to threshold value. The additional thresholds allow for the identification and analysis of the characteristics of households closer to the poverty line, and to those most exposed to the risk of falling below the line. The use of additional thresholds is particularly useful where the threshold value used is sensitive to peaks in the income distribution introduced by the government transfer system. For example, in Australia, there are peaks in the income distribution which reflect the rate of various income support payments and if a threshold value is positioned close to one of these peaks, it is possible that in a given year the proportion of persons above or below the line could vary depending on relative changes in the value of those payments. The application of different poverty thresholds in EU‐SILC is described in Box 3.14.
Recommendation 14: For international comparisons of relative poverty among CES countries, a 50% threshold is recommended for the main indicator, in order to ensure consistency with the global SDG 10 indicator on relative poverty. In addition to the global SDG indicator, this threshold is also consistent with that used by the OECD in reporting on relative poverty in member states.
This measure may be complemented by the use of additional thresholds (such as 60% in EU countries), in order to provide further context when comparing across groups of countries, particularly relating to sensitivity of poverty rates to the choice of threshold.
Box3.14 EU‐SILCuseofdifferentrelativepovertythresholds
European at‐risk‐of‐ (relative) poverty levels using a threshold line set at the 70% of the median value are about four times higher than those that obtain when using a 40% threshold. However, the relative rankings of individual countries can change significantly with uses of different thresholds. For example, Malta shows the second lowest at‐risk‐of‐poverty rate when the threshold is set at 40% of the median—but falls to 19th when the threshold is set at 70% of the median (the rate rises from 2.4% to 25.7%). This means that a significant share of the population (almost 25%) has equivalised incomes very close to 50%
Chapter3MonetaryPoverty
76
of the median value. The opposite case is apparent in Romania, where the at‐risk‐of‐poverty rate using the 70% threshold is “only” 2.3 times greater than that which obtains with the 40% line. The share of the population with equivalised incomes close to half of the median value is about 17%.
Table 1 Individual at‐risk‐of poverty thresholds and rates for national populations, 2014
Poverty threshold ((OECD‐modified) equivalised disposable incomes in Euros) set at:
At risk of poverty rate
(share in total population)
40% median
50% median
60% median
70% median
40% median
50% median
60% median
70% median
Belgium 8,682 10,852 13,023 15,193 3.8 8.6 15.5 24.9 Bulgaria 1,324 1,655 1,987 2,318 10.8 15.9 21.8 28.0 Czechia 3,049 3,811 4,573 5,336 2.4 5.2 9.7 17.0 Denmark 11,144 13,931 16,717 19,503 4.4 6.6 12.1 20.4
Germany 7,893 9,867 11,840 13,813 5.4 10.5 16.7 23.7
Estonia 2,887 3,609 4,330 5,052 7.9 13 21.8 29.0 Ireland 8,068 10,084 12,101 14,118 4.3 8.8 16.4 24.4 Greece 3,072 3,840 4,608 5,376 10.4 15.8 22.1 28.9 Spain 5,308 6,634 7,961 9,288 10.6 15.9 22.2 29.7 France 8,480 10,600 12,719 14,839 2.9 6.7 13.3 21.9 Croatia 2,090 2,613 3,135 3,658 8.2 13.4 19.4 27.0 Italy 6,303 7,879 9,455 11,031 8.7 12.7 19.4 26.6 Cyprus 5,760 7,200 8,640 10,080 3.3 7.8 14.4 24.2 Latvia 2,081 2,601 3,122 3,642 7.9 13.2 21.2 29.2 Lithuania 1,929 2,411 2,894 3,376 6.9 11.3 19.1 26.6 Luxembourg 13,728 17,160 20,592 24,024 4.0 8.1 16.4 24.4 Hungary 1,805 2,256 2,707 3,159 4.5 9.2 15.0 22.4 Malta 5,115 6,394 7,672 8,951 2.4 8.4 15.9 25.7 Netherlands 8,356 10,446 12,535 14,624 2.8 5.9 11.6 19.2 Austria 9,284 11,605 13,926 16,247 4.0 8.2 14.1 21.2 Poland 2,135 2,668 3,202 3,735 5.8 10.7 17.0 24.8 Portugal 3,291 4,114 4,937 5,760 8.6 13.8 19.5 27.1 Romania 862 1,077 1,293 1,508 13.1 19.1 25.1 31.0 Slovenia 4,764 5,955 7,146 8,337 4.1 9.1 14.5 21.6 Slovakia 2,724 3,405 4,086 4,767 5.1 8.4 12.6 19.7 Finland 9,481 11,851 14,221 16,591 2.5 5.5 12.8 22.2 Sweden 10,848 13,560 16,272 18,984 4.7 8.5 15.1 22.5 United Kingdom 8,211 10,264 12,317 14,369 5.0 9.5 16.8 25.6 Iceland 8,995 11,243 13,492 15,741 2.1 3.9 7.9 15.6 Norway 17,510 21,887 26,265 30,642 3.9 6.2 10.9 17.7
Source: Data from EU‐SILC surveys ilc_li01 and ilc_li02. Available from http://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=ilc_li01 and http://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=ilc_li02.
Chapter3MonetaryPoverty
77
3.3.3.3 Anchoredpovertylines
228. Anchored poverty lines are sometimes used to supplement more “standard” relative poverty measures, as they bring some of the strengths of absolute poverty measures while also being considerably more straightforward to use.
229. An example is Eurostat’s at‐risk‐of poverty rate anchored in time. The measure is obtained using the at‐risk‐of‐poverty threshold in a particular year, adjusted for inflation during the following years. Comparison of changes in this measure with those in the “standard” at‐risk‐of‐poverty rate gives an indication of changes in the absolute situation of those with low incomes in relation to changes in the relative situation. In other words, the former takes explicit account of changes in overall price levels, so if there is an increase in real incomes (as typically there is) it implies that everyone, including those at risk of poverty, becomes better off over time. In contrast, the standard measure accounts for changes in average income levels (including the price effect and changes in real income).
230. By comparing the results obtained using the anchored poverty line with the standard one (60% of the median), it is possible to evaluate the differences.
Figure3.2 At‐risk‐of‐povertyratesanchoredatafixedmomentintime(2008)versus“standard”at‐risk‐of‐povertyrates,2008‐2014
Source: Eurostat.
231. In Figure 3.2, for the EU as a whole, we observe that during 2008‐2011 the anchored at‐risk‐of‐poverty rate is slightly lower than the standard rate—indicating that living standards for the population increased a little more than prices. In the following years the situation reverses: the at‐risk‐of‐poverty rate that obtains with the anchored line is higher than what obtains with the standard rate. Due to the economic crisis, growth in median incomes was lower than the price increase. Moreover, because the anchored measure is adjusted only for inflation, the anchored at‐risk‐of‐poverty rate can be interpreted as the share of the population who can afford to purchase a fixed (in 2008) basket of goods and services. However, the composition of this basket is not really identified; nor is it possible to
5.0
10.0
15.0
20.0
25.0
30.0
2008 2009 2010 2011 2012 2013 2014
European Union (27 countries) anchored
European Union (27 countries)
Cyprus anchored
Cyprus
Slovakia anchored
Slovakia
update availabl
3.3.3.4
232. Wcountrysame rause of internaof averthoughtthe povconsumnationacountribecomecountri
3.3.4
3.3.4.1
233. Ais typicathe indi
234. Indifferenare in pthe houmonetaeconomsmallerunit eqgoods eliving toliving aincrease
235. W(KlasenZaidi, 20
its value byle for vulne
4 Weak
With a stany grow at tate. Ravallioa “weakl
tionally, therage nationt to be metverty threshmption per cal poverty les. If rich ce 1 (i.e. thes).
Keyissu
1 Equiv
s highlighteally the houividual.
n light of tnt sizes andpoverty. Theusehold is tary costs omies of scaler budget thaquivalenciesenjoy econoogether canalone (econes).
While there, 2000), the002):
One relyingapproach);
One using d
y taking intorable house
klyrelativ
dard relativhe same raon and Chenly relative”e line is connal consumt. Above thhold could capita, withlines). This countries whe ‘weakly
es
valencesc
ed above, thusehold or
his, it is esd compositioe simplest ao use a perof satisfyinge in consuman adults tos). Additionomies of scan cover thenomies of s
e is no genere are at l
g on beha
direct quest
Chapter
o account peholds.
vepoverty
ve povertyate. Similarn (2009) arg” poverty nstant (e.g.mption (e.g.is level, in take the fo
h a gradientapproach
were considerelative’ li
calesand
he unit of ofamily, whi
ssential thaon be placealternative fr capita povg an individmption. Thiso satisfy thenally, multiale with regir needs wiscale or de
nerally accleast three
vioural ana
tions to obt
r3Monetary
78
price trends
yline
y line, povely, poverty gue that thiline. With, PPP$1.25/., PPP$2 a light of theorm of a lit of a third was develoered, it migne becomi
economie
observationile the unit
at individuaed on an eqfor linking tverty line. Hdual’s needs runs couneir food andple individgard to heatithout needecreasing m
cepted metmain appro
alysis to es
tain subject
yPoverty
for new go
erty will nowould notis is implaush such a /day in 200day) whe
e growing imne that inc(a value esoped basedght be expeng ‘fully re
esofscale
for incomeof analysis
als living inual footing he value ofHowever, thds are consnter to the ed clothing nuals living ting and houding to spenmarginal co
thod for caoaches tha
stimate equ
ive estimat
oods and se
ot fall when rise if all isible and arline for m05 prices) ure absolutemportance creases in sstablished bd on data fected that telative’ for
e
e or consums for povert
n householwhen assef the poverthis implicitlystant, and evidence theeds (i.e., ttogether ausing. As a nd twice asost when t
alculating et are often
uivalence s
es (subjecti
rvices that
n all incomincomes fergue insteadmeasuring p to a certae basic neof social exslope with based on dafor 115 devthe gradienvery high
mption expey should id
ds (or famssing whethty line to thy assumes tthat there hat childrenhere are coand sharingresult, two s much as ahe househ
equivalence used (Dea
scales (beh
ive approac
become
mes in a ll at the d for the poverty ain level eds are xclusion, average ata from veloping nt would income
enditure deally be
milies) of her they e size of that the are no
n need a onsumer g public persons a person old size
e scales ton and
avioural
ch);
Chapter3MonetaryPoverty
79
One that simply sets scales in some reasonable, but essentially arbitrary, way (arbitrary or parametric approach).
236. The first two methods are, for conceptual and econometrical reasons, not fully convincing (Deaton and Zaidi, 2002; Deaton, 1997). Most studies to date are therefore based on arbitrary (or parametric) equivalence scales.
237. One of the most widely used equivalence scales (based on the parametric approach) is the so‐called “OECD‐modified scale”, which assigns a value of 1 to the household head, of 0.5 to each additional adult member, and of 0.3 to each child. This scale was adopted by Eurostat in the late 1990s.
238. The square root scale is also commonly used in the CES region. Recent OECD publications comparing income inequality and poverty across countries use a scale that divides household income by the square root of household size. This implies that, for instance, a household of four persons has needs twice as large as one composed of a single person. However, some OECD country reviews, especially for Non‐Member Economies, apply equivalence scales that are in use in their respective country.
239. At a national level, a variety of practices are used, including country specific equivalence scales. The use of equivalence scales in Polish and Russian poverty measures are described in Box 3.15 and Box 3.16.
240. The choice of a particular equivalence scale depends on technical assumptions about economies of scale in consumption as well as on value judgements about the priority assigned to the needs of different individuals such as children or the elderly. These judgements will affect results. For example, the poverty rate of the elderly will be lower (and that of children higher) when using scales that give greater weight to each additional household member, since children tend to live in larger households than do the elderly (Förster, 1994). In selecting a particular equivalence scale, it is therefore important to be aware of its potential effect on the level of inequality and poverty, on the size of the poor population and its composition, and on the ranking of countries. The impact of the parameters used to assess the relative poverty risk among children and those over 65 is shown in Table 3.3.
Table3.3 Relativepovertyriskofchildrenvstheelderly,bytypeofequivalencescale
Country Year Parameter of equivalence scale
theta=1 theta=0.75 theta=0.5
Albania 1996 3.5 3.5 1.6
Armenia 1998/99 1.3 1.0 0.7
Azerbaijan 1999 1.1 1.0 1.0
Belarus 1999 2.0 0.9 0.4
Bulgaria 1997 1.5 0.9 0.6
Croatia 1998 0.9 0.4 0.2
Czechia 1996 40.3 21.3 1.1
Georgia 1996/97 1.5 0.9 0.7
Hungary 1997 7.4 2.9 1.0
Kazakhstan 1996 1.4 1.0 0.7
Chapter3MonetaryPoverty
80
Country Year Parameter of equivalence scale
theta=1 theta=0.75 theta=0.5
Kyrgyzstan 1997 1.2 1.1 0.9
Latvia 1997/98 2.7 1.6 0.8
Lithuania 1999 2.3 1.4 0.7
TFYR Macedonia 1996 1.7 1.4 0.9
Republic of Moldova 1997 1.6 1.0 0.7
Romania 1998 4.2 2.0 0.6
Russian Federation 1998 1.2 0.9 0.7
Slovenia 1997/1998 1.0 0.6 0.3
Tajikistan 1999 1.2 1.2 1.0
Turkmenistan 1998 1.8 1.7 1.3
Ukraine 1996 1.1 0.7 0.4
Source: Calculations based on World Bank (2000). Note: Relative poverty risk of Children vs Elderly shows ratio of child (0‐15) to elderly (65+) poverty
risk, i.e. 1 means equal risk of poverty, 2 means children poverty risk is twice higher than elderly, while 0.5 means elderly poverty is twice higher. Theta is a parameter of simplified one‐parameter equivalence scale, Equivalent size = (household size)θ. OECD scale is close to θ=0.5, while θ=1 implies no economies of scale, and θ = 0.75 is a reasonable estimate for transition countries used by the World Bank (2000).
241. The impact of applying different scales is explored in more detail in the UNECE Canberra Handbook (2011).
242. The adjustment of household income through equivalisation allows for more accurate comparative analysis between incomes of families of different sizes and composition, and it is recommended particularly for the analysis of relative poverty in international (or interregional) comparisons, as well as within countries over time.
Recommendation 15: In setting a poverty line, equivalised welfare measures should be used. For international comparisons, a trade‐off needs to be made between applying country‐specific approaches reflecting variation in economies of scale and ensuring comparability across the region. For comparisons across the CES region, it is recommended that the square root scale be used in order to provide consistency with existing international statistics based on a 50% of median threshold contained within the OECD income distribution database for many CES countries.
Use of an alternative scale (such as modified‐OECD) for this headline measure could lead to inconsistencies in the levels of relative poverty reported by the UNECE and OECD, reducing the coherence of international statistics. However, use of such alternative scales would be useful for any supplementary measures to streamline comparison among a set of countries that are relatively homogenous (from a global perspective).
Chapter3MonetaryPoverty
81
Box3.15 InfluenceofdifferentequivalencescalesonpovertyratesinPoland
The influence of different equivalence scales on poverty rates (head count ratio) was first analysed by Poland’s Central Statistical Office on the basis of Poland’s 2013 EU‐SILC survey. The definition of poverty adopted by Eurostat was applied. According to this definition, people are considered poor if they lived in households falling below 60% of the median equivalised disposable income of the total population.
Table 1 Relative poverty rates in Poland calculated with different equivalence scales
Poverty threshold
Equivalence scale
Persons aged
Total 0‐17 18‐65 65 or more
% of persons in households
60% of median disposable income
Original OECD equivalence scale 1‐0.7‐0.5 18.0 26.3 17.5 9.5
Modified OECD equivalence scale 1‐0.5‐0.3 17.3 23.2 16.7 12.3
Square root scale sqrt(n) 17.7 22.6 16.7 15.8
Per person n 20.1 33.2 19.2 7.7
The results of the analysis indicated that the use of different equivalence scales in Poland does not have a significant impact on poverty estimates at the level of the whole population. However, it does influence the results broken down into different age groups. For the whole population, the lowest value of the relative poverty rate (approximately 17%) was generated by the modified OECD equivalence scale. The highest value (20%) came when the effects of the household’s structure were omitted.
As mentioned, the effects of applying different equivalence scales were more apparent when the analysis was conducted at the level of different age groups rather than the whole population. Generally, it can be said that assigning a higher weight to each additional household member (in particular to children) results in higher poverty rates among the youngest persons and lower rates among the oldest ones.
For example, the relative poverty indicator for Polish children (persons aged 0‐17) assessed via the original OECD equivalence scale was estimated at approximately 26%. When calculated according either to the modified OECD equivalence scale or the square root scale were, it was some 3 percentage points lower. Likewise, when relative poverty rates for the elderly (65 and above) were calculated on a basis of the original OECD equivalence scale, approximately every tenth person could be considered as being at risk of poverty. Applying the modified OECD equivalence scale increased the at‐risk‐of‐poverty rate for this group of persons to 12%; in the case of the square root scale, it reached almost 16%.
These differences underscore how different methodological choices (including the choice of equivalence scale) can have a significant influence on reported poverty rates, as well as the identification of the social groups that can be considered the most vulnerable to poverty. This is important to remember in the context of policies that seek to reduce poverty.
Chapter3MonetaryPoverty
82
Box3.16 EquivalisationinRussianpovertymeasurement
Economies of scale resulting from cohabitation (everything else equal) occur for reasons related to sharing of certain costs, in particular related to payments for housing and communal services, purchases of vehicles or newspapers, household appliances, etc.
In Russia, a 1996 Russian State Statistics Committee (Rosstat) study of primary household budget survey microdata found that the savings achieved from cohabitation in households surveyed did not exceed 5% total living costs. The absence of substantial empirical confirmation of the effect of cohabitation can be explained by the fact that about 50% of consumer spending in low‐income households is spent on food, while non‐food expenditures on goods and services relate mainly to personal consumption. In other words, the basic expenses in poor households are of an individual nature, and cannot be consumed jointly.
Using equivalisation scales for determining absolute poverty:
Under these conditions, the use of statistically unsuitable equivalence scales can lead to artificially low levels of absolute poverty (see Tables 1 and 2). The magnitude of absolute poverty depends directly on the equivalence scale chosen (or equivalency ratio E) and can, all else equal, differ at times.
For example, as shown by experimental calculations carried out on the basis of the 2014 population income survey (with 2013 data), the absolute poverty indicator, calculated by Rosstat without equivalisation (Е=1) was 11.1%. The application of the equivalency ratio E = 0.73 reduced the value of the poverty rate to 5.0%, and at E = 0.5 ‐ to 2.7%.
Table 1 shows the values of absolute poverty levels, and Table 2 shows the age structure of the absolutely poor population for different equivalency ratios (in percentages).
Table 1 Values of absolute poverty levels
Total Younger than Working
age Older than working age working age
Е=1 11.1 20.5 10.9 3.2
Е=0.73 5.0 9.3 5.1 1.2
Е=0.5 2.7 4.7 2.8 0.6
Table 2 Age structure of the absolutely poor population for different equivalency ratios (Percentages)
Total Younger than Working
age Older than working age working age
Е=1 100 35.9 57.9 6.2
Е=0.73 100 35.8 59.2 5.0
Е=0.5 100 33.9 61.5 4.6
Note: In calculating absolute poverty, Rosstat does not use an equivalence scale, since the value of the subsistence minimum (absolute poverty line) for a household is generally defined in terms of its composition as a sum of relevant indicators set out in the specific constituent entity of the Russian Federation for different socio‐demographic groups, taking into account a calculation of basic expenses for personal consumption.
Chapter3MonetaryPoverty
83
Using equivalisation scales in determining relative poverty:
While the application of equivalisation scales has only a slight impact on the at‐risk‐of‐(relative) poverty rate for the population as a whole, it can have a more significant impact on the reported composition of the poor.
For the experimental calculation of the relative poverty of the general population conducted on the basis of 2014 population income survey in 2014, a poverty line of 50% of the median per capita income level was used, and three values of the coefficient of equivalence (Е = 1, 0.73, 0.5) were examined.
Table 3 shows relative poverty levels and Table 4 shows the structure of the relatively poor population by main age groups depending on the equivalency ratio (in percentages).
Table 3 Relative poverty levels
Total Younger than Working
age Older than working age working age
Е=1 15.6 26.7 14.8 8.0
Е=0.73 15.1 22.9 13.6 12.0
Е=0.5 15.7 20.2 13.1 18.8
Table 4 Age structure of the relatively poor population for different equivalency ratios (Percentages)
Total Younger than Working
age Older than working age working age
Е=1 100 33.2 55.7 11.1
Е=0.73 100 29.5 53.1 17.3
Е=0.5 100 25.0 49.1 25.9
3.3.4.2 PricesandPPPs:TheInternationalComparisonProgramme
243. Cross‐country comparisons of poverty rates depend crucially on information about prices in various countries (except where fully relative measures of poverty are used). This information allows researchers to compare welfare between individuals living in different countries by adjusting domestic incomes by purchasing power parity (PPP) exchange rates, so that one international dollar affords, in principle, the same command over goods and services in any country of the world.29 PPP exchange rates play a role similar to that played
29 For example, if the price of a hamburger in France is €4.80 and in the United States it is $4.00, the PPP for hamburgers between the two economies is $0.83 to the euro from the French perspective (4.00/4.80) and €1.20 to the dollar from the US perspective (4.80/4.00). In other words, for every euro spent on hamburgers in France, $0.83 would have to be spent in the United States to obtain the same quantity and quality—that is, the same volume—of hamburgers. Conversely, for every dollar spent on hamburgers in the United States, €1.20 would have to be spent in France to obtain the same volume of hamburgers. To compare the volumes of
Chapter3MonetaryPoverty
84
by national price indexes in the case of individual countries over time. In order to compare average or individual welfare in the same country in two periods, one needs to adjust for changes in national price levels. Similarly, to compare welfare between individuals living in different countries at the same point in time, one needs an estimate of price levels they face. Cross‐country comparisons of absolute poverty rates are thus sensitive to estimates of PPP exchange rates.
244. These estimates are obtained through a large International Comparison Programme (ICP).30 The ICP is a joint project of the United Nations, OECD, the World Bank and regional development banks, which measures national price levels. The project entails direct comparisons of prices for some 1000 goods and services to construct country‐wide price indexes for total GDP and such broad components as household consumption, investment, and government spending, and for narrower components of expenditures on goods like clothing and footwear, and transport. These comparisons are carried out at approximately decennial intervals.
245. PPPs are relative price ratios that are calculated in several stages: first for individual goods and services, then for groups of products, and finally for each of the various levels of aggregation adding up to GDP. In moving up the aggregation hierarchy, the relative price ratios refer to increasingly complex assortments of goods and services. Thus, if the PPP for GDP between France and the United States is €0.95 to the dollar, it can be inferred that for every dollar spent on GDP in the United States, €0.95 would have to be spent in France to purchase the same volume of goods and services.
246. Purchasing the same volume of goods and services does not mean that the baskets of goods and services purchased in both economies will be identical. The composition of the baskets will vary to reflect differences in tastes, cultures, climates, price structures, product availability, and income levels. But both baskets will, in principle, provide equivalent satisfaction or utility. PPP indexes are further standardised by expressing them in a common currency unit. The common currency used for the global comparison is the US dollar, and so each economy’s PPP is standardised by dividing it by that economy’s dollar exchange rate. The standardised indexes so obtained are called price level indexes (PLIs or $PPP).31
247. Since the early 1990s, the World Bank has monitored global extreme poverty using an international poverty line that was explicitly based on the national poverty lines of some of the world’s poorest countries. Each release of new PPP data has led both to revisions of the international poverty line and to re‐assessments of the relative differences in well‐being across countries and regions.
hamburgers purchased in the two economies, either the expenditure on hamburgers in France can be expressed in dollars by dividing by 1.20 or the expenditure on hamburgers in the United States can be expressed in euros by dividing by 0.83. 30 For more details on the ICP, see http://go.worldbank.org/X3R0INNH80. 31 Economies with PLIs greater than 100 have price levels that are higher than that of the base economy. Economies with PLIs less than 100 have price levels that are lower than that of the base economy. So, returning to the hamburger example, if the exchange rate is $1.00 to €0.79, the PLI for a hamburger with the United States as the base economy is 152 (1.20/0.79 × 100). From this, it can be inferred that, given the relative purchasing power of the dollar and the euro, hamburgers cost 52% more in France than they do in the United States.
248. Tothree stin natioparticulgeneratshare ohousehrevisionsince thpoint ge
249. Thlines wain the 1updatednew seglobal p
250. Thresults the newaggregascholar2005 seline.
251. Inpovertyand inf2011 psimple which tinterna
252. It rich andpatternspend aaccountthe cou
3.4 P
3.4.1
253. Hnext sta
o measure teps are unonal currenlar ICP rounte poverty of the popold incomen, poverty rhe beginninenerally pro
he first inteas set at $11990 World d this to $1et of nationpoverty line
he latest ICwas presenw estimatesates for all ts, the domiet, creating
n 2015 they lines for 1lating themrices, theseaverage wathe World tional pove
is importad poor withns of poor a large prot for a relatuntry as a w
Poverty
Overview
aving decidage is the s
poverty indertaken. Fncies at thnd. Secondlines in napulation live or expenrates are reng of measoduced is 19
ernational 1.01 using 1Developme1.08 per danal poverty e of $1.25 pe
CP round wnted to the ps of PPP exthe participinant view the need f
e World Ba15 very poom to 2011 ue national las taken. ThBank rounderty line.
nt to note hin economhouseholdsportion (oftively small hole.
indicato
w
ded on a wselection of
Chapter
n different First, the inte benchma, this poveational currving belownditure surcalculated surement o981).
poverty line985 PPPs, bent Report ay, using 19lines, Ravaer day.
was conducpublic. The change ratepating counis that thesfor another
ank revisedor countriesusing each cines were he result of ded up to $
that PPPs omies. This ms differ systften a majoproportion
ors
welfare meaf one or mo
r3Monetary
85
countries ternational ark year urty line is arencies bacw this poveveys. It is not only fof poverty a
e that was by Ravallion(World Ban993 PPPs. Wallion, Chen
cted in 201new estimaes, and acctries. Thougse new PPPr revision to
d its interns (expressedcountry’s oconverted those oper$1.90, and
offer compamay turn outematically ority) of then of the spe
sure and eore indicato
yPoverty
using thesepoverty linsing the Padjusted uskwards anderty line isimportant r the most at the glob
based on n, Datt and nk, 1990). CWith the ren and Sang
11, and in Oates of priceordingly negh there waPs represento the World
ational povd in local cwn consuminto dollarsrations yieldwhich repr
arisons acrout to be profrom those
eir incomesending of th
stablished ors useful fo
e internatione is turnedPP exchanging nationad forward s determinto empharecent yeaal level (w
a sample ovan de Wahen and Ralease of thraula (2009
October 20e levels in 1ew $PPP estas some disted an impd Bank’s int
verty line urrency un
mer price ins using theded $1.88 present the
oss economoblematic, e of the be on basic she better‐of
at least onor tackling
onal povert into a povege rates fral inflation in time. Thned using asise that, rs, but for ahere the fi
of national lle (1991) aavallion (200e 2005 PPP9) proposed
014, the fu199 countrietimates of sagreementrovement oternational
by taking its at 2005ndex. Then, e 2011 PPPsper person new World
mies, not acsince the setter‐off. Tstaple foodsff, and ther
e poverty lpoverty. In
ty lines, erty line rom the rates to hird, the national in each all years rst data
poverty nd used 01) later Ps and a d a new
ll set of es led to national t among over the poverty
national prices), once in s, and a per day, d Bank’s
ross the pending he poor s, which efore of
line, the dicators
may bepoverty
254. Thcountripublish poverty
255. Mstatic mmeasurtransitiogiving aand grinterve
3.4.2
3.4.2.1
256. Thproportexpendand megroups.
257. DFirst, wdepth odistinctsignificaare stilldepth o
258. Thpolicymwould bwho are
e used to hy that peopl
he measurees and inte a suite of y.
Monetary pomeasures bres that maons into ana headline inroups, dynntions to ta
Staticme
1 Headc
he most ction of the iture is lesseasure, allo. This can be
1
10 NP
N
i
where 0P i
(or sample)
is less than
espite its swhile it descof poverty ttion is madeantly belowl below theof peoples’
his featuremaking. If thbe to focuse very poor
ighlight thele experienc
es describedernational omeasures,
overty indicased on incake use of nd out of pondication ofamic measackle povert
easures
countrati
ommonly upopulation
s than the powing userse expressed
)( zyI i
is the propo
) and I (‐)
the povert
strengths ancribes the nthat people e between w. One imple poverty lipoverty inc
e can alsoe focus is sos on those g, which wou
Chapter
e level of poce, and how
d below havorganisatiowhich allow
ators can bcome or colongitudinoverty. Brof current levsures are ty effectivel
io
used measn that is livpoverty lines to easily ud as:
ortion of th
is a functio
y line ( z )
nd ubiquitynumber of experiencethose who ication of tne), there reases, the
o potentialolely on thegroups whould arguably
r3Monetary
86
overty in dw poverty is
ve strengthns tend now those us
broadly be gonsumptional data to oadly speakvels of poveof more ly.
ure is the ving in houe. It is popuunderstand
he populatio
n that take
and 0 if iy
y, the headcpeople whe. It is basedare just behis is that iwill be no indicator a
lly lead toe headcouno are just by be more s
yPoverty
ifferent cous changing o
hs and weakot to focus sing the dat
grouped inton at a givenconsider ping, while serty and house in he
headcounuseholds whlar because the scale
on that is p
s a value of
is greater t
count ratioo are in pod on a binaelow the pof poor indivchange in lso will not
o perverset ratio, the elow the psocially ben
untries or aover time.
knesses. Foron a singleta a more r
o two categn point in tpoverty ovestatic measw they varyelping polic
t ratio, whhose income it is easy tof poverty
oor, N is th
f 1 if income
than z .
has a numoverty, it dory measureoverty line viduals becthe indicatbe affected
incentiveeasiest wayoverty line,neficial.
areas, the d
r that reasoe indicatorrounded pi
gories of metime; and der time, as ures are usy across placy makers
hich descrime or consuto both undamongst d
he total pop
e/expenditu
mber of limoes not refe of povertyand those wome less poor. Similarld.
s with regy to reduce , rather tha
depth of
on, most , but to cture of
easures: dynamic well as seful for ce, time design
bes the umption derstand different
pulation
ure (iy )
itations. flect the y and no who are oor (but y, if the
gard to poverty an those
Chapter3MonetaryPoverty
87
3.4.2.2 Povertygapindex
259. The poverty gap index measures the extent to which individuals fall below the poverty line (the poverty gaps) as a percentage of the poverty line. The poverty gap index can be expressed as:
N
i
i
z
G
NP
10
1
where the poverty gap (
iG ) is equal to the value of the poverty line less actual
(equivalised) income or expenditure for individuals in poverty, and zero for those who are not in poverty.
260. The sum of these poverty gaps can be seen as the minimum cost of eliminating poverty, if it were somehow possible to perfectly target social transfers.
261. The division by the poverty line normalises the measure, allowing for comparisons across countries and across time.
262. The poverty gap ratio also has its limitations, however. In particular, the measure only reflects the average depth of poverty, so cannot reflect changes in inequality among the poor. Additionally, it can actually rise rather than fall when people leave poverty, if the average poverty gap of those that remain increases as a result. An additional consideration is that data on the very lowest incomes can often be affected by poor data quality, which in turn will affect the usefulness of poverty gap measures.
263. An alternative method of providing a sense of the depth of poverty is to examine headcount ratios using lower thresholds. For example, in Canada, very few seniors are found below 30% of the low income measure (LIM), indicating that "depth of poverty" is less severe for this group (because of guaranteed income supplements for low income seniors).
3.4.2.3 Squaredpovertygap
264. The squared poverty gap index averages the squares of the poverty gaps relative to the poverty line. This implicitly puts more weight on observations that are well below the poverty line, thereby taking into account inequality among the poor. However, the squaring of the poverty gaps means that it is less easy to interpret than the standard poverty gap index.
265. It is one of a class of poverty measures proposed by Foster, Greer and Thorbecke (1984) which vary the weight of the income (or expenditure) level of the poorest members in society. These measures are additively decomposable. They also allow separating changes into a component resulting from rising average incomes and a component resulting from changes in the distribution of income. The use of these measures in Russian poverty statistics is illustrated in Box 3.17.
Chapter3MonetaryPoverty
88
Box3.17 PovertyindicatorsinRussianFederation
Possibilities for calculating absolute poverty indicators in Russian Federation are determined in part by the availability of relevant information at the time of compiling the statistics. The production of statistics for each reporting period is carried out in several stages, the results of which constitute preliminary and final estimates of the indicators. The details of the choice of income measure and of the indicator of the level of absolute poverty at the preliminary and final stages are described below:
1. At the preliminary assessment stage:
Criterion of income—monetary income of population (macro assessment): This includes employee wages and salaries (based on payroll data and adjusted for arrears), earnings of persons engaged in entrepreneurial activities, pensions, allowances, scholarships and other social transfers, income from property, interest on deposits, securities, dividends, and other income. The calculation of monetary income for the population includes an adjustment for hidden compensation, which is defined as the difference between total household expenses (including the growth of their financial assets) and officially registered income.
Absolute poverty indicator: The number of people with incomes below the subsistence minimum (headcount measure) is calculated based on use of analytical models in accordance with the procedures approved by Rosstat in 1996, by agreement with a number of interested ministries and agencies.
The share of the population with incomes below the subsistence minimum is calculated by the following lognormal formula:
.0,
2
1)(
;0,0
;; 2ln0
2
xdteuF
x
xzLu t
x
where x
xzu
ln
0lnln
; 2
xln0 5,0lnxln ;
μ – macro‐value of per capita income;
xln ‐ average quadratic deviation of income logarithms determined on the basis of the empirical distribution of population income according to the results of Population Income Survey; z – subsistence minimum in the average per capita.
Note: A similar approach is used in calculating the MDG indicator “Proportion of population whose dietary energy consumption is below the minimum allowed level”. In determining the proportion of people whose dietary energy consumption is below the minimum level, a logarithmic function is used.
2. At the stage of final assessment:
Criterion of income—monetary income of population (measured using the population income survey): This includes labour incomes (the sum of remuneration before payment of income tax, including the monetary value of benefits provided by employers, at the main place of employment, income from self‐employment, including gross income from sales of
Chapter3MonetaryPoverty
89
products (services) of own production, income from other labour activities, in addition to the main job); income from interest earned on savings; income from rental property; income from the lease (sublease) of land; and transfers received (social benefits, including pensions, benefits, compensation and other social benefits; cash receipts from individuals and organisations other than the social security authorities, including child support and other payments equal to them).
Absolute poverty indicator: the share of the population with incomes below the subsistence minimum is estimated on the basis of the survey data comparing the income of each household surveyed with the value of the subsistence minimum as determined by household composition (as the sum of the relevant figures set out in the specific constituent entity of the Russian Federation for the different socio‐demographic groups). Estimated shares of the population with incomes below the subsistence minimum are produced by the formula:
n
i
i
z
xz
nР
1
0
0 0;max1
where
z –subsistence minimum in the average per household member;
xi ‐ per capita income index value of i‐person surveyed;
n ‐ total number of population surveyed.
The poverty gap ratio (Р1), which characterizes the average distance of poor people from the poverty line, is also calculated, using the formula:
n
i
i
z
xz
nР
1
1
1 0;max1
The poverty severity ratio (Р2), which characterizes the degree of inequality among poor people, is calculated by the formula:
n
i
i
z
xz
nР
1
2
2 0;max1
The difference between the poverty gap ratio and poverty severity ratio is that by it is calculating a greater weight given to households with a significant lack of funds.
Indicators Р0, Р1 and Р2 combined into a class of poverty by Foster, Greer, and Thorbecke:
n
i
i
z
xz
nР
1
0;max1
Chapter3MonetaryPoverty
90
3.4.2.4 Person‐equivalentpoverty
266. Despite the importance of tracking changes in the depth of poverty, measures such as the poverty gap index have had relatively limited use in policy formation and monitoring due to being deemed “unintuitive” and difficult to understand.
267. The person‐equivalent approach, developed by Castleman, Foster and Smith (2015), seeks to address this problem, whilst keeping the desirable characteristics of poverty gap measures. Person‐equivalent headcount measures benchmark the initial conditions of the poor, with this benchmark then being used to sum the number of person‐equivalents to get a headcount measure. Someone who is twice as far below the poverty line as a standardised person is counted as two person‐equivalents, whilst someone who is only half as poor would be counted as half a person‐equivalent.
3.4.2.5 Othermeasures
268. There are a number of other static measures with characteristics that make them desirable as indicators. However, they lack the intuitive appeal of the straightforward measures presented above. A short overview of the Watts index and the Sen‐Shorrocks‐Thon index is given below.
269. The Watts index divides the poverty line by income, takes logarithms, and finds the average over the poor. The use of logarithms means that, as with the squared poverty gap, the Watts index is much more sensitive to changes in the lowest incomes than it is to changes for those with higher incomes. It is also possible to decompose the measure by group or region.
270. The Sen‐Shorrocks‐Thon index was developed from the now relatively little used Sen index. It is the product of the headcount index, the poverty gap index and a term that uses the Gini coefficient of the poverty gap ratio.
)ˆ1(10PP
SST GPPP
271. One of its key strengths is the possibility for decomposition, allowing users to understand whether changes in the overall poverty index are being driven by changes in the number of people who are below the poverty line, the depth of that poverty, or the level of inequality amongst the poor population.
Recommendation 16: For regional poverty measures, it is recommended that the primary indicator is the headcount ratio, due to its widespread acceptance in policy and ease of comprehension. Poverty data producers should consider the value of adopting other indicators, such as the poverty gap ratio or person‐equivalent poverty at the national level. The headcount ratio should be reported alongside the value of the poverty line for a single adult household (in PPP).
3.4.3
272. Aprovide
3.4.3.1
273. It detrimeof stratsuch asrisk of simpact adverseincreasPopli, 2povertywhich cexperiemeasurindicatoCommispopulatrisk‐of‐pyears. Ifollowe
274. Boand ent
Dynamic
nalysing poed by static
1 Persis
is widely aental for thtegies to de reducing esocial excluof persist
ely affectinging the like2014). In addy decrease tcan make usence lengthres of persors in use.ssion, whiction living inpoverty thrts calculatied over four
ox 3.18 protry and exit
cmeasur
overty trendmeasures.
stentpov
acknowledge individuaeal with shoexpenditureusion for thtent povertg their coglihood that dition, Fouathe longer se of longituhy spells ofsistent pove. Perhaps h defines tn householreshold for on requiresr years.
ovides examrates in the
Chapter
res
ds can prov
erty
ged that expl than a briort‐term droe or making ose who brty on childnitive devethey will earge and Laan individuudinal data f poverty aerty. Therethe most the persisteds where ththe currens a longitud
mples of thee United Kin
r3Monetary
91
ide an impo
periencing ef period inops in incomuse of saviriefly fall indren in paelopment, pxperience pyte (2005) al remains to help ideare invaluae are a nuwidely useent at‐risk‐ohe equivalist year and dinal instru
e analysis ongdom and
yPoverty
ortant addit
poverty ovn poverty. Ame, which dngs or loanto poverty.rticular caparticularly poverty as ahave shownin poverty.
entify those able to polmber of vaed one is of‐poverty rsed disposaat least twment, thro
of both persother EU co
tion to the
er a numbeA householdo not appls. These str. Studies han be espein the firstadults (see n that the c. For these groups thalicy makersariants of that used rate as the able incomewo out of thugh which
sistent at‐riountries.
information
er of years d can use aly in the lonrategies redave shown tecially detrt years of le.g., Dickerchances of ereasons, int are more s. One exapersistent by the Eupercentage
e was belowhe precedinthe individ
sk‐of‐pover
n that is
is more a variety ng term, duce the that the imental, life, and rson and escaping dicators likely to ample is poverty uropean e of the w the at‐ng three uals are
rty rates
Box3.1Persist
Figure 1rate forwere atthan hacomparlowest poverty
Figure 1Persiste
S
were notavailable
18 tentpover
1 shows bor EU countrt persistent alf the overison with olevels of py out of the
1 ent and sing
Source: OfficeNote: Persistet available at for Sweden,
rtyintheU
oth the overries in 2013risk of poverall relativother individpersistent p28 membe
gle year at‐
e for National ent poverty ratime of pubIreland, or Cro
Chapter
UnitedKin
rall at‐risk‐o3 (ONS, 201erty, equivave at‐risk‐odual EU coupoverty (i), er states (ii).
‐risk‐of‐pov
Statistics, Eurates are the lalication, so 20oatia. Overall
r3Monetary
92
ngdomand
of‐poverty 15). In 2013alent to appf‐poverty runtries, in 2but had t
.
verty rates f
rostat. atest available012 figures wpoverty estim
yPoverty
dEU
rate and th3, 7.8% of pproximatelyrate, which2013, the Uthe 13th hi
for EU coun
e. For Bulgariwere used. Nomates are all 2
he persistenpeople in ty 4.6 millionh in 2013 sUnited Kingdighest leve
ntries, 2013
a, Romania ao persistent p2013 rates.
nt‐at‐risk‐ofhe United Kn people. Thstood at 1dom had onl of cross‐s
3
nd Greece 20poverty estim
f‐poverty Kingdom his is less 5.9%. In ne of the sectional
013 figures ates were
This rewhen cbelow. at leastindicatione of tof 58%.of the othe vastBy contlikely to
Figure 2Rate of
S
were the
lationship bconsidering A ratio of 5t two out ong that lessthe lowest . In contrasoverall povet majority otrast, in theo be for a sh
2 f persistent
Source: OfficeNote: For Bule latest availab
between pthe ratio b
50% would sof the last ts than half oof the EU ct to the Unerty rate foof people ine United Kihorter perio
poverty as
e for National lgaria, Romanble for persist
Chapter
ersistent abetween thsuggest thathree yearsof those in ountries fonited Kingdor 2013; in It relative incngdom, forod.
s a percenta
Statistics, Eurnia, and Greecent poverty in
r3Monetary
93
nd overall he two ratet half of thos. In 2013, poverty thar which datom, the pertaly it was 6come poverr those exp
age of overa
rostat. ce, the ratio in these count
yPoverty
poverty raes expresseose currentthe Unitedat year had ta are availarsistent pov69%. This surty experieneriencing r
all relative
s calculated uries
tes can bed as a percly in povertd Kingdom been persiable and beverty rate inuggests thanced it overelative low
poverty acr
using 2012 po
e most cleacentage in ty were alsohad a ratiostently poolow the EUn Romania wt in these cr a number income, it
ross the EU
overty rates s
arly seen Figure 2 o poor in o of 49% or. This is average was 81% countries of years. is more
, 2013
ince these
275. Bothe Eur
Box3.1Povert
Betweeinto popovertyCommis(Figure
Figure 1Rates o
The firsSpain, bin a rela
The seLithuanpovertypoint of
In the tCzechiashare ogreater
ox 3.19 proopean Com
19 tyentryan
en 2008 andoverty fromy in 2008 ssion, 20121).
1 of entry and
st group of but also incatively posit
cond grounia, Hungaryy, and low f view, as it
hird group,a, Finland, aof persistenr churning in
ovides an exmmission.
ndexitrate
d 2009, som one year had manag2). However
d exit from
countries, ludes to a ltive situatio
p of couny, Malta, Pochances of reflects a h
low risks ond the Netnt poor is hn Cyprus, De
Chapter
xample of a
esinEUco
me 6% of thto the nexged to exir, the entry
risk of pove
which is messer extenon where bo
ntries (consortugal, Romf escaping phigh risk of
of entering iherlands, thhigh compaenmark, Lux
r3Monetary
94
nalysis of p
ountries
e EU populxt, while 40it from poy and exit ra
erty, 18‐64
ost clearly nt Belgium, oth entry an
sisting of mania, and Ppoverty. Thbeing trapp
into povertyhis turns ouared to thexembourg,
yPoverty
poverty entr
lation as a w0% of the poverty by tates varied
year olds
representeFrance, Irelnd exit rate
Bulgaria, EPoland) shohis situationped in pover
y are combut to be a sige risk of poSlovenia an
ry and exit
whole was population the followinconsiderab
ed by the Uland, Austris are high.
Estonia, Greows both a hn is problemrty.
ined with logn of social overty. In cnd Sweden.
rates condu
likely to hathat was ang year (Ebly across c
United Kingdia, and Slov
eece, Italyhigh risk of matic from
ow exit ratepolarisatiocontrast, th
ucted by
ve fallen at risk of uropean countries
dom and akia, are
, Latvia, entering a policy
es. In the n, as the here is a
3.4.3.2
276. Ainto anlimited
277. Thwere nothe exityear am
278. Siexit ratrates aseach ca
RecommtargetinCES regcountrilongitudproduce
3.5 Imin
3.5.1
279. OAgendagoals athe glob
280. Pomanifesinequalsecurity
281. Thfully suand uppthe poorelevanless rel
32 This is concepts
2 Entry
nother impd out of popanel dura
he entry raot in povertt rate is defmong those
ince there aes expresses a percentase would e
mendation ng policy efgion overales producidinal data, e comparab
mprovinndicato
Poverty
On 25 Septea for Sustainnd targets bal, regiona
overty redustations by ities withiny.
he Millenniitable for cper‐incomeor live on wnce for theievant to th
true where ths or countries.
yandExit
portant appoverty fromtions make
ate into povty one yearfined as thewho were a
are fewer ped as a perctage of thoquate to a m
17: Dynamffectively. Al due to thing EU‐SILCfrom eithe
ble dynamic
nginterrsandm
andtheS
ember 2015nable Develare designeal, and natio
uction is ce2030. Addit and among
um Develocomparing pe countries, well above r stage of dhe UNECE r
here are more.
Chapter
rates
plication of m one year analysis of
verty is genr earlier bute percentagat‐risk‐of‐po
people in pocentage of ose not in pmuch larger
mic measureAlthough nohe limited aC data, NSer survey oc poverty in
rnationametadat
SDGs
5, the Unitopment Goed to be reonal levels.
entral to thetionally, pog countries,
pment Goapoverty in twhere phy$1 per daydevelopmenregion as a
e people out o
r3Monetary
95
longitudinato the nexpoverty du
nerally meat fell into pge of individoverty the y
overty thanthose in popoverty. Smr percentag
es of povero dynamic iavailability SIs should or administdicators in
alcompta
ted Nationoals (SDGs) levant for t
e SDGs. Goverty dimen, and in Goa
al (MDG, 20the UNECE ysical survivy. Many Mnt. Similarlywhole. Ho
of poverty tha
yPoverty
al data is txt. This canuration chal
asured as thpoverty in thduals not atyear before
n not in povoverty will umall changege change fo
rty are a vandicators aof suitable consider
trative southe future.
parabilit
s General with 17 glothe whole w
oal 1 calls fonsions are pal 2 on endi
000‐2015) inregion. All
val is generaDG indicatoy, some of wever, they
n in. This may
he examinan be particulenging.
he percentahe followin‐risk‐of‐pove.
verty, it is tousually32 bes in the nuor those in p
aluable toolre currentlylongitudinaopportunitrces, in or
ty:Regio
Assembly bal goals anworld and t
or an end tpresent in Ging hunger
ndicators onUNECE coually given, aors were ththe global y may be s
y not be the c
ation of traularly usefu
age of peog year. Converty in the
o be expecte higher thaumber of pepoverty.
in developy proposedal data apaties for prrder to be
onal
adopted thnd 169 targto be monit
to poverty Goal 10 on rand achiev
n poverty wuntries are and the maherefore ofSDG indicasupplement
ase for all gro
ansitions ul where
ple who nversely, current
ted that an entry eople in
ping and d for the art from roducing able to
he 2030 gets. The tored at
in all its reducing ing food
were not middle‐ jority of f limited tors are ted with
oups,
additionthe guidthe SDG
Recommto the S
3.5.2
3.5.2.1
282. Gdisaggr
283. Recountri
284. Racountriapproppovertythat an set out
3.5.2.2
285. Gdisaggr
286. Re
287. Rathe wea
288. Dand sexchapteremploy
nal indicatode thereforG targets.
mendation SDG targets
Monetar
1 TargeeveryPPP$1
lobal indicegated by s
egional indes.
ationale: Tes. Furtherriate for aly levels are absolute pin Section 3
2 Targewomedimen
lobal indicaegated by s
egional ind
ationale: Thak internati
isaggregatiox (as per tr, it is also ryment statu
ors that arere sets out a
18: Monet.
rypovert
et1.1:Byywhere,cu1.90/day
ator: 1.1.1sex and age
dicator: No
The PPP$1.rmore, detll CES counconsidered
poverty line 3 of this cha
et1.2:Byenandchnsionsacc
ator: 1.2.1 sex and age
icator: Prop
he global inonal compa
on: As a mithe global recommends, househo
Chapter
e monitoreda regional a
tary poverty
tyindicat
2030,eraurrentlym(in2011
Proportiogroup and
specific re
.90/day extermining antries, is nod useful acrfor that grapter and in
2030,redildrenofcordingto
Proportiongroup.
portion of p
dicator is earability of n
inimum, theindicator). ded that furld type, and
r3Monetary
96
d at the napproach to
y indicators
torsforC
adicateexmeasuredprices).
n of popuemployme
ecommenda
xtreme pova single abot feasible. ross groupsoup of coun more deta
duceatleaallageslionationa
n of popula
opulation li
qually apprnational po
e regional iIn line witrther disaggd urban/rur
yPoverty
tional and monetary p
s for the UN
CEScount
xtremepodaspeopl
lation belont status.
ation for a
verty line bsolute poWhere dir
s of similar ntries be esail by Jolliffe
astbyhalivinginpoaldefinitio
ation living
iving below
ropriate for verty lines
ndicator shth the recogregation bral classifica
regional levpoverty ind
NECE region
tries
overtyforlelivingo
ow internat
absolute po
is not apoverty line,rect compacountries, stablished ue and Prydz
lfthepropovertyinons
below nat
w national po
CES countrand definiti
hould be disommendatibe provided tions.
vel. This seicators, in l
n should be
allpeoplonlesstha
tional pove
overty line
propriate which woarisons of ait is recomusing the pr (2016).
portionoallits
tional pove
overty line.
ries, acknowions.
saggregatedions earlierwhere feas
ection of ine with
aligned
lean
erty line
for CES
for CES ould be absolute mended rinciples
fmen,
rty line,
wledging
d by age r in this sible, by
3.5.2.3
289. Gdisaggr
290. Redisaggr
291. RanumberIncomefeasiblestatistic
292. Dand sexalso recstatus,
3.5.3
293. ATerms povertyto explinforma
294. InAt the iprices. (2005) alongsidthe bas
295. MmonitoIn sommethod(Unitedspecific
296. Innot cleareferenimmineprovide
3 Targeandprace,e
lobal indicaegated by a
egional indegated by a
ationale: Thr of countre Distributioe, to avoid cs. The OEC
isaggregatiox, as the glocommendedhousehold
Metadat
t the natiosuch as “y” should beain how naation correc
n measuringnternationaIn the pasthave beende the dataket used fo
Maintenancering MDGs e cases, thdology is nd Nations, cation make
n general, dar to which ce to the pent importaed in nation
et10.2:Bypoliticalinethnicity,
ator: 10.2.1age and sex
dicator: Prage and sex
he global inies in the ron Databasa proliferaD definition
on: As a mibal indicatod that furthtype and ur
aconside
onal level, p“relative poe used in coational povctly.
g poverty, eal level, thet, PPP$1.00/ used. Thea, even if stor the PPP co
e of good mhas revealehe exact oot clearly s2008). Pu
es comparat
data sourceindicator oprimary souance that sal reports.
Chapter
y2030,emnclusiono,origin,re
1 Proportio.
oportion o.
dicator is eegion will ae (IDD), it tion of simns are in line
inimum, theor. In line wer disaggrerban/rural c
erations
poverty linoverty”, “aonjunction werty lines w
even variatie standard m/day (in 19ese changesandard defonversion, c
metadata is ed that metor even brospecified anublishing ttive analysis
s are listedor which peurces. Thesuch metad
r3Monetary
97
mpoweraofall,irreeligionor
on of peop
of people
qually appralready be makes sen
milar indicate with the r
e regional iwith the recoegations be classificatio
es are set bsolute powith their swere deter
on in the bmeasure of 985 prices), s show theinitions arecan be cove
important ftadata wereoad definitnd sometimthe nations difficult.
d in nationariod they ree are oftenata be ava
yPoverty
andpromspectiveoreconomi
ple living be
living belo
ropriate for reporting ose to applytors and resrecommend
ndicator shommendatiprovided wns.
using variooverty”, “sespecific defimined, ens
ase year caPPP$1.90/dPPP$1.08/
e importance used. In reered.
for cross‐coe often misions are mmes even tnal poverty
al reports. Hefer. Importn missing oilable along
otethesoofage,sexicorothe
elow 50%
ow 50% o
CES countron this indicy the samesulting condations set
hould be disons earlier
where feasib
ous definitievere poveinitions. Mesuring users
an make resday is used,day (1993) ce of includeports, furth
ountry compsing, incommissing. In he definitioy line wit
However, intant for ther not specig with the
ocial,econx,disabilierstatus
of median
of median
ries. As a sigcator for the definitionsfusion for out in this c
saggregatedin this chapble, by emp
ions and merty” or “eetadata are s can interp
sults incom, measuredand PPP$1ding the mher details,
parability. Wmplete, or inseveral cason is not phout any
n several cae interpretafic enoughpoverty es
nomicity,
income
income
gnificant he OECD s where users of chapter.
d by age pter, it is loyment
methods. extreme needed pret the
parable. in 2011 1.25/day metadata such as
Work on ncorrect. ses, the provided further
ases it is tion is a . It is of stimates
Chapter3MonetaryPoverty
98
297. At the international level, both the OECD and Eurostat set out metadata reporting requirements for countries contributing to the IDD and producing EU‐SILC data respectively. UNECE has also developed a publication, “Getting the Facts Right” (2013), which provides a guide to presenting metadata, with particular emphasis on the MDGs.
Recommendation 19: Metadata are important for helping users understand the extent to which figures are comparable across countries and over time. This is particularly the case where indicators are based on national poverty lines, which allows for considerable variation in practice between countries. For monetary poverty indicators for the UNECE region, it is recommended that the following minimum set of metadata be made available, in order to assist users in making sensible comparisons both between countries and within countries over time:
Conceptual metadata
Unit of observation (e.g., household) Unit of analysis (e.g., individual) Population covered (e.g., private households) Definition of welfare measure, including information on any deviation from the
main international standards (e.g., UNECE Canberra Handbook (2011)) Equivalence scale used: (e.g., square‐root scale) Type of poverty line: Absolute or relative (for indicators based on national poverty
lines) Methodology for calculating poverty line (for indicators based on national poverty
lines) Reference period: Period of time or point in time to which the measured
observation is intended to refer Unit of measure: Unit in which the data values are measured (e.g., headcount
ratio, percentage of population).
Methodological metadata
Data provider: Organization that produced the data Source data: Characteristics and components of the raw data used for compiling
statistical aggregates, (i.e., type of primary source [e.g., survey, census, registry]) and any other relevant characteristics (e.g., sample size for survey data).
Contact information: Individual or organizational focal points for the data, including information on how to reach them (e.g., website, mail address, phone, e‐mail).
Quality metadata
Comparability: Explanations should be provided where differences between statistics can be attributed to differences between the true values of statistical characteristics. Comparability issues can be broken down into:
Geographic differences: degrees of comparability between statistics measuring the same phenomenon for different geographical areas;
Temporal differences: degrees of comparability between two or more instances of data on the same phenomenon in the same country measured
3.6 S
298. Thinternaout in measur
3.6.1
3.6.2
A
at PeriodicitTimelinesAccuracy:the statisvariance ((e.g., cove
Summar
his sectiontional compthe previorement of m
Unitofo
Recommof obserreasons.
Recommwith thepopulatio
Recommcover priextendinstudy exadata, in oIt is essstatistics
Disaggre
Recommdisaggregindicatorstatus, ho
It is furtbreakdow
ge:
0‐17 (chi
18‐24
different py: e.g., anns: Number Closeness tics are int(random ererage, samp
ryofrec
n provides parability ous sectionsmultidimens
observati
endation 1.rvation sho
endation 2e indicatorson living in
endation 3ivate houseg this coveamples in torder to estential to i.
egationo
endation 4.gate povertrs for the Uousehold ty
ther recomwns.
ldren)
Chapter
oints in timual, every fof months aof computatended to mror). This mpling, non‐r
commen
a summaof statistics s of this chsional pover
on/analy
. In produciould be th
. Poverty ss describinhouseholds
3. It is recoeholds. It israge. This mthis chaptertimate povenform use
ofdata
. Given the ty indicatorNECE regioype, disabili
mended th
r3Monetary
99
me. ive years, eafter incomations or esmeasure. Thmay be descesponse) or
ndations
ary of the on monetahapter. Recrty (chapter
ysisandp
ng data on he househo
tatistics shog, for exas below the
ognised thats recommenmay involver, or utilisinerty in popurs about t
importancers wheneven should bety status22 a
hat the foll
yPoverty
tc. me/consumpstimates to his includescribed in terr measures
s
recommery poverty commendatr 5).
populati
income or old, for bo
ould be repmple, the poverty lin
t the majonded that Ne research sng alternativulation grouhe coverag
e of disaggrer possible.e disaggregaand urban/
owing clas
ption referethe exact os bias (systrms of majoof accuracy
endations fand the retions 1‐4 a
oncover
consumptiooth practica
ported at tnumber oe.
rity of povNSIs explorsuch as thave data souups that arege of the
egation, it i. As a miniated by agerural popul
sifications
ence periodor true valuematic erroor sources oy.
for improvlated metadalso pertain
rage
on, the noral and con
he individuof individua
verty statistre the feasiat given in turces, inclue difficult tpublished
is recommeimum, the e, sex, empation.
be used fo
. es that or) and of error
ving the data set n to the
mal unit nceptual
ual level, als in a
tics only ibility of the case ding big o reach. poverty
ended to poverty loyment
or these
Em
H
U1.2.3.
3.6.3
25‐49
50‐64
65 and ov
mployment
Employed
Unemplo
Retired
Other ou
Household ty
One‐per
Two adu
Two adu
Two adu
One adu
Other
Urban/rural2
. Predomi
. Intermed
. Predomi
Welfare
Recommenbe the maactual incosaving. Howcountry orsupplemen
Recommenbased on co
Recommenstrengths aavailability.a given poapproachesregion, it iwidespreadother areas
Recommentogether, itpoverty stexpenditurtechniques
ver
t status23:
d
oyed
tside the la
ype:
rson househ
ult househo
ult househo
ult househo
ult househo
24: inantly urbadiate regioninantly rura
measure
ndation 5. Itin income ome that inwever, to r area, comtary income
ndation 6. Wonsumption
ndation 7. and weakne. Where botopulation, ts. Howeveris recommed usage ams of the regi
ndation 8. Gt is recomatistics cone as well awhere pos
Chapter
abour force
holds;
ld without c
ld with one
ld with two
lds with chi
an region n al region
es
t is recommmeasure undividuals wprovide ad
mpilers of e measures
Where consn expenditu
Both incomesses as povth income athere is var, for internended thatong EU andion.
Given the admended thnsider examas their intsible.
r3Monetary
100
children;
e child unde
o or more ch
ildren unde
mended thatsed for povwithin a hodditional inpoverty stas, such as in
sumption is ure.
me and coverty measuand consumalue in utilnational comt income bd OECD cou
dvantages ohat, where mining poversection, t
yPoverty
er 18;
hildren und
r 18;
t annual (eqverty measousehold hasights intoatistics maycome befor
used as a w
nsumption ures. Their cmption expeising povermparisons be the mainuntries as w
of consideridata availaverty meastaking adva
er 18;
quivalised) surement, aave availab the natury also wishre social tra
welfare me
expendituchoice shouenditure darty measurof poverty n welfare well as incre
ing multipleability allowsures basedantage of s
disposableas this refleble for spenre of poveh to makeansfers.
easure, it sh
re have pauld depend ta are availes based oacross themeasure, gasing availa
e welfare mws it, compd on incomstatistical m
income ects the nding or rty in a use of
hould be
articular on data lable for on both e UNECE given its ability in
measures pilers of me and matching
3.6.4
Recommenservices froacross counpoverty inpurposes, supplemenownership internationnational andevelop neand inequa
Recommenthe value ooperationasame reasexpenditurfor the purp
Recommencommon printernationstatistical cincome aninternationon data in guidance forganisatio
Recommenincome or poverty onstatistics thaim of beinwealth, in This shouldavailable toshould be through the
Poverty
Recommenpurposes, itmore stableof the distr
Recommencountries, ensure con
ndation 9. om owner‐ntries, it is ndicators ucompilers tary measin other wanal comparind internatiw guidelinelity statistic
ndation 10. of services l definitionon, they ae in practicpose of inte
ndation 11. ractice, it isnal poverty compilers cnd consumnal best prawhich the
for nationaons.
ndation 12. consumpti its own. Ithat can be ang able to order to prd be possibo producerutilised whe same surv
line
ndation 13. t is recomme thresholdibution
ndation 14. a 50% thresistency wi
Chapter
Due to th‐occupied drecommen
used for inof pover
sures incluays, such asison in futuonal level, es on the mcs.
In practiceby househo of incomeare also exce. It is theernationally
As accounts recommencomparisoconsider demption expectices, so te effects ofal statistica
While weon in asset is recommassessed aloconsider jorovide a coble when rers of statishere incomvey source.
In setting mended thatd which is le
For internaeshold is rith the glob
r3Monetary
101
he challengdwellings anded that snternationarty statistiding impus using an afure, as weit is recomeasuremen
e, because old consum set out in xcluded frorefore recoy comparab
ing for the nded that thns (at leasteveloping menditure sthat future if these tranal offices s
alth is an ssing povermended thaongside othointly the dmplete pictegisters andstics. Alternme (or cons
relative pot the mediaess affected
ational comrecommendbal SDG 10
yPoverty
ges associatnd the varuch serviceal compariics may fted rent, fter housingell as the tmended thnt of impute
of the chaler durablesthe Canbeom the mommended le poverty s
value of sohey be exclt for now). methods fortatistics, ainternationnsfers are should be
important rty, it cannat countriesher welfare istribution ture of indid other adnatively, stasumption) a
overty lines an is used a by measur
mparisons oded for theindicator o
ted with mriation in mes be excludson. Howefind it usor take ag costs meatargeting ofhat internated rent for i
llenges invos, they are rra Handboeasurementhat the sastatistics.
cial transfeuded from It is also rr including nd invest al comparisincluded. Tdeveloped
factor to not be useds invest in measures, of income, ividuals’ ecoministrativeatistical maand wealth
for internas a paramerement issu
of relative pe main indin relative p
measuring methods emded from thever, for seful to caccount ofasure. To bef resourcesional organnclusion in
olved in meexcluded fook (2011). nt of consuame practic
rs in kind isindicators urecommendthese tranin learnin
sons may bTo assist wd by inter
consider ald as a meadevelopingwith the loconsumptionomic wee data souratching tech are not a
ational comter, as it pres towards
poverty amoicator, in opoverty. In a
housing mployed he main national consider f home etter aid s at the nisations poverty
easuring rom the For the umption ce apply
s not yet used for ded that nsfers in ng from be based with this, national
longside asure of g wealth ng‐term ion, and ll‐being. rces are chniques available
mparison ovides a the top
ong CES order to addition
3.6.5
3.6.6
3.6.7
to the globOECD in recomplemenorder to pparticularly
Recommenbe used. Fapplying coensuring corecommendwith existinwithin the alternative inconsistenreducing thscales wouamong a se
Indicato
Recommenprimary indand ease oadopting opoverty at value of the
Recommenand targetproposed flongitudinaopportunitisources, in the future.
Regional
Recommenaligned to t
Metadat
Recommenextent to particularlyallows for c
bal SDG indeporting onnted by theprovide fury relating to
ndation 15. For internatountry‐specomparabilityded that thng internatiOECD incoscale (such
ncies in thehe coherencld be usefuet of countr
rs
ndation 16. dicator is thof comprehother indicathe nationae poverty lin
ndation 17. ting policy for the CEal data aparies for prodorder to b
lpoverty
ndation 18. the SDGs ta
a
ndation 19.which daty the case considerabl
Chapter
dicator, thisn relative pe use of addrther conteo sensitivity
In setting ational comcific approay across thehe square rional statistme distribuh as modifi levels of rce of internul for any sies that are
For regionhe headcouension. Povators, suchal level. Thene for a sing
Dynamic meffectively
ES region ort from couducing longbe able to p
ymeasur
Monetary rgets.
Metadataa are comwhere indie variation
r3Monetary
102
threshold poverty in ditional threext when cof poverty
a poverty liparisons, aches reflece region. Foroot scale btics based oution databied‐OECD) frelative povnational stasupplement relatively h
nal poverty nt ratio, duverty data h as the pe headcoungle adult ho
measures of y. Althoughoverall duentries prodgitudinal daproduce co
res
poverty ind
are impormparable accators are between co
yPoverty
is also conmember sesholds (sucomparing rates to the
ne, equivala trade‐off ting variatior comparisobe used in on a 50% obase for mafor this heaverty reporttistics. Howtary measuhomogenou
measures,ue to its widproducers poverty gapnt ratio shoousehold (in
poverty areh no dynae to the limucing EU‐SIata, from eiomparable d
dicators for
rtant for hcross countbased on ountries.
sistent withstates. Thisch as 60% iacross gro
e choice of t
ised welfarneeds to on in econoons across torder to pf median thany CES couadline meated by the wever, use res to streaus (from a g
it is recomdespread acshould conp ratio or uld be repon PPP).
e a valuablemic indicatmited avaiILC data, NSither surveydynamic po
r the UNEC
helping usetries and onational po
h that useds measure in EU countoups of cothreshold.
re measuresbe made bomies of scthe CES regprovide conhreshold countries. Ussure could UNECE andof such alteamline comglobal persp
mmended tcceptance insider the vperson‐eq
orted along
e tool in devtors are clability of SIs should cy or adminoverty indic
E region sh
rs understaover time. overty lines
d by the may be tries), in ountries,
s should between cale and gion, it is sistency ontained se of an lead to d OECD, ernative mparison pective).
that the in policy value of uivalent side the
veloping currently suitable consider istrative cators in
hould be
and the This is
s, which
Chapter3MonetaryPoverty
103
For monetary poverty indicators for the UNECE region, it is recommended that the following minimum set of metadata be made available, in order to assist users in making sensible comparisons both between countries and within countries over time:
Conceptual metadata
Unit of observation (e.g., household) Unit of analysis (e.g., individual) Population covered (e.g., private households) Definition of welfare measure, including information on any deviation from main international standards (e.g., UNECE Canberra Handbook (2011)) Equivalence scale used (e.g., square‐root scale) Type of poverty line: Absolute or relative (for indicators based on national poverty lines) Methodology for calculating poverty line (for indicators based on national poverty lines) Reference period: Period of time or point in time to which the measured observation refers Unit of measure: Unit in which the data values are measured (e.g., headcount ratio, percentage of population).
Methodological metadata
Data provider: Organization that produced the data. Source data: Characteristics and components of the raw statistical data used for compiling statistical aggregates, i.e., type of primary source (e.g., survey, census, registry) and any relevant characteristics (e.g., sample size for survey data). Contact information: Individual or organizational focal points for the data, including information on how to reach them (e.g., website, mail address, phone, e‐mail).
Quality metadata
Comparability: Explanations should be provided where differences between statistics can be attributed to differences between the true values of statistical characteristics. Comparability issues can be broken down into:
Geographic differences: degrees of comparability between statistics measuring the same phenomenon for different geographical areas; Temporal differences: degrees degree of comparability between two
or more instances of data on the same phenomenon measured at different points in time.
Periodicity: e.g., annual, every five years, etc. Timeliness: Number of months after income/consumption reference period. Accuracy: Closeness of computations or estimates to the exact or true values that the statistics are intended to measure. This includes bias (systematic error) and variance (random error). This may be described in terms of major sources of error (e.g., coverage, sampling, non‐response) or measures of accuracy.
Chapter4PovertyDashboardsandtheMaterialDeprivationIndices
104
4 PovertyDashboardsandtheMaterialDeprivationIndices
4.1 Introduction
299. This chapter explains the measurement of non‐monetary aspects of poverty and social exclusion. Reducing such deprivations is essential to meeting the Sustainable Development Goals (SDGs). Measuring non‐monetary deprivations is part of poverty measurement because the SDGs clearly regard poverty as multidimensional. The SDGs focus on reducing poverty “in all its forms and dimensions”. Some national and regional policies already address non‐monetary deprivations in such areas as housing, health, education, and other services. The chapter shows how countries can introduce a basic dashboard of social indicators and indices of material deprivation.
4.2 Processesandprinciples
300. The development of poverty indicators has often been the responsibility of national governments. More recently, some indicators have been estimated by international organisations. For example, the World Bank generated the PPP$1/day index in 1990, and has developed subsequent methodological revisions until its current form of PPP$1.90/day. UNICEF, UNESCO and other agencies, together with data providers using demographic and health surveys, also contributed to the standardisation of data and indicators in such areas as malnutrition, education, and health.
301. The Millennium Development Goals (MDGs) saw the development of relevant metadata and data quality standards for harmonised statistics by the Inter‐Agency Expert Working Group on the MDG indicators and the United Nations Statistical Commission (UNSC). However, this process was criticised as being too centralized. The UNSC‐led process of developing comparable SDG indicators has been considerably more inclusive, involving consultation with the 28 government members of the Inter‐Agency Expert Group but also observer governments, UN agencies, academics, and civil society actors, inter alia via online open consultations. What seems clear is that the procedural aspects of indicator development and approval are vital and cannot be overlooked.
302. Drawing on such experiences as the EU’s open method of coordination (OMC) in developing the EU‐SILC surveys and common measurement standards, Tony Atkinson and Eric Marlier (2010) proposed to UNDESA principles and processes for the development of comparable indicators of poverty and social exclusion. This section summarizes their recommendations, which would be relevant insofar as UNECE develops harmonized datasets.
303. In terms of procedures, Atkinson and Marlier “draw on our experience from the construction of social indicators in the European Union and in their actual use in the policy process (Atkinson et al., 2002; Marlier et al., 2007), because there are, in our view, lessons to be learned about the way EU member states cooperate through the OMC. The OMC
Chapter4PovertyDashboardsandtheMaterialDeprivationIndices
105
process has limitations, but it illustrates concretely how 27 countries can reach agreement on common objectives and monitoring procedures and how evidence‐based policymaking can be aided by comparative analysis and international benchmarking. … The fight against poverty and social exclusion is a common challenge, and there is scope for mutual learning, despite the differences in circumstances and in levels of living” (2010:387).
304. Atkinson and Marlier (2010: 45) outline five criteria for internationally comparable indicators of deprivation in social inclusion:
305. In October 2016, the World Bank launched the Atkinson Commission Report “Monitoring Global Poverty” (World Bank, 2017). This report highlights the importance of multidimensional poverty, and proposes that indicators be collected universally, to create both a dashboard of social indicators and a multidimensional poverty index (MPI) that reflects overlap between component indicators. It suggested (like this chapter) that the national policy be given priority, but that social indicators be designed with an eye towards obtaining at least partial comparability across countries because of the rich analysis that this can support. The Atkinson Report also proposed that MPI design follow due process and involves all nations and stakeholders. An interesting example of such a process is the OMC followed in the European Union (Box 4.1).
Box4.1 EU‐SILCandtheOpenMethodofCoordination
A key feature of EU‐SILC is the process by which it was developed: the Open Method of Coordination. This method balanced national priorities with progressive harmonisation of data and targets.
“The open method of coordination, which is designed to help member states progressively to develop their own policies, involves fixing guidelines for the Union, establishing quantitative and qualitative indicators to be applied in each member state, and periodic monitoring” (Atkinson et al. 2002, 1–5).
EU‐SILC is replete with interesting lessons. For example, many surveys are only representative at the national level, but some sample sizes are much larger. Certain
An indicator should identify the essence of the problem
and have an agreed normative interpretation.
An indicator should be robust and statistically
validated.
An indicator should be interpretable in an
international context.
An indicator should reflect the direction of change and be susceptible to revision as improved methods become
available.
The measurement of an indicator should not impose
too large a burden on countries, on enterprises, or
on citizens.
questiocomparaddressEU‐SILCmonito
EU‐SILCThe hepoverty
306. Aappropstatic oby the p
307. Invisible s
a) b)
c)
4.3 C
308. A each ofand setkey ind
Recommprioritie
4.3.1
309. Prof an agdevelopgood ex
310. Dharmondefinitiotabulat
Ch
ons (e.g., lere across cs. Also, the C project, aring.
C data haveadline targy and social
tkinson andriate indicaor dynamic, public, issue
n addition, Asocial indica
The portfolThe indicatshould be pThe portfocitizens
Compara
poverty daf them usingt of prioritieicators.
mendation es, and to n
Requirem
rocess: Thegreed set ofped a set ofxample of s
Data sourcenise survey ons of grouion, and rep
hapter4Pove
evels of edountries (Ause of admand studies
been usedget of reduexclusion h
d Marlier (2ators shouldinput‐ , oues such as e
Atkinson anators:
lio of indicators shouldproportionalio of indic
abledas
ashboard shg just one ines, and pote
20: It is dame a sma
mentsfo
e developmf indicatorsf key indicauch a proce
es: Followiquestions ups by whicporting form
ertyDashboa
ucation, seAlkire and Aministrative s have sho
d since 2010cing (by 20has been de
2010: 8–14) d be stock otput‐ , or oease of inter
nd Marlier p
ators should be mutuaate cators shou
shboard
hows levelsndicator. It entially to e
esirable foll set of 5‐1
rcompar
ent of coms and definittors based ess.
ng agreemor administch indicatormats.
ardsandtheM
106
elf‐reportedApablaza, 2data alongsown both i
0 to monito0 million inefined based
also providor flow, suboutcome‐barpretation a
propose thr
d be balancelly consiste
ld be as tr
ds
s of deprivais desirableeven name
r dashboar0 indicators
rabledas
mparable dations. Europon nationa
ment of intrative or rrs are to be
MaterialDepr
d health st2016)—an iside survey ts potentia
or poverty an 2020) thed on the EU
de an overvbjective or osed, and soalso affect i
ree principle
ed across thnt and the
ransparent
ation in diffe for dashboa small set
rds to haves as tier 1 o
shboards
shboards repe’s open ml action pla
dicator deregistry date disaggreg
rivationIndic
atus) may ssue that fdata has bal and diffi
and social ee numbers U‐SILC instru
view of the objective, reo on. Whenndicator se
es for the s
he differentir weight w
and access
ferent dimeoards to hav of 5‐10 ind
e a clear hier key indica
s
equires firsmethod of cans and furt
efinitions, ca sources, gated, perio
ces
still be diffuture surveen exploreculties for
exclusion inof people ument.
purposes foelative or an statistics alection and
selection of
dimensionwithin the p
sible as pos
ensions, preve a clear hdicators as t
erarchy andators.
st the identcoordinationther discuss
countries nsample desodicity, met
fficult to eys may ed in the poverty
n the EU. living in
or which bsolute, are used design.
a set of
s portfolio
ssible to
esenting ierarchy tier 1 or
d set of
ification n, which sion, is a
need to sign and thods of
311. Rprovide
4.3.2
312. Thapproacthe 18 tevaluatPPP$1.2rate, thprovidedimens
313. This ongoand civdevelop
4.3.3
314. Boexclusioin the R
Box4.2Europe
The opeinclusioand soc
a) b) c) d) e)
Table 1social e
Ch
eporting: Aed by a coor
ExampleSDGs
he MDGs cach—they wtargets to ated with in25/day, thehe share ofed a multiions and tra
he developing at the tvil society pment of M
Example
ox 4.2 shoon and poveRussian Fede
2 eansocial
en methodon and socicial exclusio
OverarchinSocial incluPensions Health careInvesting in
1 below shoexclusion tar
hapter4Pove
A comparabrdinating in
esofcom
an serve as were a dashachieve the ndependente share of uf parliameni‐faceted packed chang
ment of thetime of writactors, th
MDG indicato
eofdashb
ws the exaerty. Box 4.eration.
indicators
of coordinal protectioon target:
g portfolio sion
e and long‐tn children
ows the diffrget.
ertyDashboa
ble dashboastitution.
mparable
a prominenboard of 49eight goalst indicatorsunderweighntary seats profile of ages in these
e more thanting. Respoe process ors.
boardsin
amples of 3 represent
s
nation prodon. These i
term care
ferent indic
ardsandtheM
107
ard could b
internati
nt example 9 indicators. Improvems such as ht children uheld by wa populatioe over time
n 231 SDG nding to vehas been
nEurope
European Sts the basic
uced the Endicators c
ators used
MaterialDepr
be reported
ionaldas
of the imps initially, wments in diffthe proporunder 5 yeawomen, andon’s achiev.
indicators aery strong dconsiderab
e
Social Indicc list of dep
EU Social Incover sever
to monitor
rivationIndic
d on a com
shboards
lementatiowhich were dferent aspecrtion of pears of age, d so on. Thvements a
and their asdemands bybly more i
cators usedrivation intr
dicators in al areas Eu
r the Europe
ces
mmon platf
s:MDGsa
n of the dadefined to cts of povereople livingthe child mhe MDG inacross a ra
ssociated my both goveinclusive th
to monitoroduced by
the areas urope 2020
e 2020 pov
form, or
and
shboard monitor rty were g below mortality dicators ange of
metadata ernment han the
or social y Rosstat
of social poverty
verty and
Chapter4PovertyDashboardsandtheMaterialDeprivationIndices
108
Table 1 Indicators used to monitor the Europe 2020 poverty and social exclusion target
Indicator Definition
At risk of poverty or social exclusion rate (headline indicator for Europe 2020)
The sum of persons who are at‐risk‐of‐poverty, or severely materially deprived, or living in households with very low work intensity, as a share of the total population
At‐risk‐of poverty rate
Share of persons aged 0+ with an equivalised disposable income below 60% of the national equivalised median income. The equivalised median income is defined as the household's total disposable income divided by its “equivalent size”, to take account of the size and composition of the household, and is attributed to each household member. Equivalisation is made on the basis of the OECD modified scale.
Population living in very low work intensity (quasi‐jobless) households
People aged 0‐59, living in households where working‐age adults (18‐59) worked less than 20% of their total work potential during the past year.
Severe material deprivation rate
Share of population living in households suffering
deprivation in at least 4 of the following nine33 areas: i)
unable to pay rent or utility bills, ii) unable to keep home adequately warm, iii) unable to cover unexpected expenses, iv) unable to eat meat, fish or a protein equivalent every second day, v) does not take a week holiday away from home, or could not afford such (even if wanted to), vi) unable to afford a car, vii) unable to afford a washing machine, viii) unable to afford a colour TV, or ix) unable to afford a telephone.
Source: EU social indicators (http://ec.europa.eu/social/)
These dashboards are often useful in terms of policy evaluation (e.g., the Europe 2020 vision), monitoring progress on poverty and social exclusion targets, assessing specific social challenges facing EU countries (e.g., through the joint assessment framework), identifying the key social trends in the EU (e.g., through the social protection performance monitor), reporting on EU social policies and adequacy in terms of child poverty and well‐being, and for analytical work on social and economic policy.
Box4.3Measu
The Ruinternaand mefocus grdevelopaspectseducati
The Russervicesindicatofrequenlist of dstatistic
Basic L
S
4.3.4
4.3.4.1
315. Dwhethepromindashbo
Ch
3 ringbasic
ussian Fedetional and Reasurementroups, 200 ped a list os of life: cuonal service
ssian Task Fs from the or, parsing ncy methoddeprivationscal office sh
List of DeprInability to af
Inability to af
Inability to af
Inability to af
Absence of a
Family lives i
Inability to pa
Absence of c
Inability to af
Inability to af(install glass,
Inability to af
Source: Rosst
Assessm
1 Advan
ashboards er to have ently, and ard is essen
hapter4Pove
deprivati
eration camRussian expt. This is a administratof characteurrent conses, and cult
Force then afocus grouactual fromd, they cames which couhortlisted ar
rivations fford a meal w
fford fruits fo
fford new clot
fford new clot
TV, fridge, an
n extremely c
ay utility bills
entral heating
fford new furn
fford emergenrenew plumb
fford necessa
at.
mentsofc
ntages
are an essdashboardto promotential in the f
ertyDashboa
onsinthe
me up withperiences offirst attemtors of houserizations osumption, hture and leis
assessed thups, to definm potential e up with auld also be re presented
with meat, chi
r all the family
thing and foot
thing for child
nd washing m
close quarters
in time
g in housing a
niture instead
ncy housing rebing, fix roof l
ry and essent
comparab
sential coms: it is whee their usefollowing th
ardsandtheM
109
RussianF
h a list of f using the dpt to harmsehold budgof familial housing ansure.
he relative dne the signpopulation dashboardmonitoredd in the tab
icken, fish twi
y members al
twear
dren as they g
achine and th
(less than 5 s
nd lack of res
d of old
epairs eaks, mend th
ial medicines
bledashb
ponent of en to highe to assess hree sets of
MaterialDepr
Federation
63 basic deprivation
monize indicget surveys poverty. Thd property
demand valnificance of n demand. d of indicato. The “list oble below
ice per week
l year round
row up
he inability to
square meters
sources to buy
he fence, pain
boards
poverty mlight a set the overalcircumstan
rivationIndic
n
deprivationn method focators acrosin urban anhe deprivaty provisions
ue for the bmaterial dAfter testinors, as well of basic de
afford them i
s per person)
y enough fuel
nt the house, e
easuremenof unidimel situation nces:
ces
ns, drawingor poverty dss regions. nd rural setttions focuss, medical
benefits, godeprivation ng the list uas a suppleprivations”
n case of nee
etc.)
t. The issuensional inof a popul
g on the definition Through tlements s on five services,
oods, and for each using the ementary that the
d
e is not dicators ation. A
Chapter4PovertyDashboardsandtheMaterialDeprivationIndices
110
316. Specialised datasets. First, the single indicators that comprise a dashboard draw on different specialised datasets, which may include survey data, administrative data, or even “big data”. A dashboard makes possible the combination of these data (as component indicators) irrespective of their intrinsic comparability (of lack thereof). The information contained in the data components can also be used to design sector‐specific policies. Thus, certain indicators, which are complex or which can only be designed using specialised surveys or sample designs, are likely to appear only in dashboards.
317. Special groups. Second, the single indicators that comprise a dashboard can refer to and analyse diverse segments of the population: children, female CEOs and politicians, youth aged 15‐24, Roma, or construction workers. For example, quality of education and skills formation could be drawn from a survey of schools and school‐going children; an employment‐related indicator could be estimated from labour force surveys; an indicator of social security could draw upon administrative records, and so on. Also, if surveys are used as the data source, each survey’s sample design must be representative for the particular groups that are of special relevance to that indicator, and these will vary across indicators.
318. Familiarity and convention. Third, conventions regarding data quality and meta‐data are already in place for many social indicators, making their computation relatively straightforward. They are also familiar, facilitating their communication. Moreover, each indicator is likely to be generated by a different expert group with specialised skills and interests in the topic.
4.3.4.2 Disadvantages
319. One potential drawback of dashboards is that they may provide too much information, risking diffuse or competing priorities. As Stiglitz, Sen and Fitoussi (2009) observed, “large and eclectic” dashboards lack a sense of priority. Furthermore, dashboards do not provide an explicit weighting across indicators. These can be ameliorated if, as Atkinson and Marlier (Atkinson et al., 2002; Marlier et al., 2007) suggest, the indicators are organised in tiers, in which the “top” tier of indicators is relatively balanced across dimensions, in which their weight is proportionate, and indicators are easy to communicate and understand. However, this rarely occurs in practice.
320. Second, because dashboards present each deprivation in isolation, and may use distinct and specialised survey instruments, they do not show overlapping or joint distribution of deprivations. Yet it is often important to know who suffers multiple simultaneous disadvantages, as these may be more deeply impoverishing than experiencing just one. These cannot be shown via a dashboard. Furthermore, in terms of policy efficacy, policies that address interconnected deprivations together, in a coordinated, multi‐sectoral or integrated approach, have been demonstrated to be more cost‐effective (UNDP, 2010). Alkire and Robles (2016) have proposed that dashboards drawing on the same survey should, at a minimum, describe this joint deprivation and have proposed graphical methods for doing so.
321. Third, dashboards do not provide a headline figure. They identify different aspects of poverty individually, but do not identify who is poor overall, based on deprivations in multiple indicators. That provides a communication challenge, because a headline could be confusing if for different indicators, “poverty has gone up, gone down, and stayed the
Chapter4PovertyDashboardsandtheMaterialDeprivationIndices
111
same” (Alkire, Foster, and Santos, 2011). While momentum may be generated from updates to monetary poverty measures, this momentum can be dissipated by the complexity of a dashboard update. The relationship between income poverty measures and other elements of a dashboard may also not be clear.
322. Fourth, the costs of designing and maintaining dashboards must be considered. Dashboard indicators may be updated with different frequencies, depending on the pace of change in an indicator. While this is appropriate, dashboard updates will be required to clarify which indicators are based on new data and which are carried over from previous updates. Yet even if each indicator is not updated each year, a large dashboard based on a diversity of specialised and possibly extensive harmonised data sources implies the need to sustain each of these data sources over time, and the cost implications of this must be considered.
Recommendation 21: The costs of data production and indicator computation for dashboard indicators should be made explicit at the time of indicator selection, and the dashboard should include indicators whose collection is financially sustainable. For example, candidate indicators should be reported together with a) the average minutes data collection takes for each unit, b) the number of units required for a national statistic (sample size), c) any special sampling issues that affect costs, and d) the frequency of updating (annual, every 5 years).
4.4 Materialdeprivationindices
323. Material deprivation indices can complement monetary poverty measures by bringing into view different but related measures of material deprivation. They have come to greater prominence in Europe because material deprivation, quasi‐joblessness and at‐risk‐of‐poverty‐and‐exclusion indicators together form the EU‐2020 poverty measure. Also, in the United Kingdom, the Index of Multiple Deprivation is used for complementary policy purposes and contains multiple deprivations including non‐material deprivations such as health and employment (see 4.4.4).
324. Material deprivation indices intend to use multiple indicators to measure a single underlying condition – material deprivation. In contrast, for multidimensional measures (or multiple deprivation measures), there is no single underlying condition. Each component indicator reflects deprivations which may differ in kind, which may be interlinked, but ‘matter’ directly. The statistical validation of these two approaches is very different.
325. Therefore, the statistical methodologies used to assess validity and reliability in material deprivation indices are distinct from methodologies used to design multidimensional poverty measures, which do not posit an underlying unidimensional concept.
Recommendation 22: Because of its extensive use, countries should include the material deprivation index (in its most recent specification) in the dashboard of comparable indicators.
Like themodify Kingdom
4.4.1
326. Adevelopof compdata soa centra
4.4.2
327. This the mpovertyEurostaafford slead anhave—ithem.
328. Thtook a selectioidentify
329. Thof the “EU matthose wspecifiedepriveitems33
33 As of eseven neshoes, inhaving toregular le
Ch
e example the mate
m, countrie
Requirem
s in the caspment of ofponent itemources, inclual platform.
Example
he most fammaterial dey (pauvreté at built uposome items n adequateit is only if
he definitiolong time bon of itemsy items that
he indicato“people at rterial deprivwho live in aed items (Ged” if they hcurrently a
a) coping b) one wec) avoidind) a meal e) keeping
end 2016, discew items: To rcluding a pairo consult anyoeisure activitie
hapter4Pove
given belorial deprivas may wish
ments
se of compfficial matems, and to uding frequ.
e:Materia
miliar mateeprivation rd’existence
on it. The mthat are co
e life. Impothey canno
on of deprivbefore the s was partlt are publicl
r is computrisk of povevation rate a householdGuio and Mhave an enfadopted wit
with unexpek’s annualg arrears (inwith meat, g the home
cussions are oeplace worn‐or of all‐weatheone; To get toes; Internet; T
ertyDashboa
ow from Reation indicto explore
parable dasrial deprivaassess and ency and d
aldepriv
rial deprivarate. Eurose) for Europmaterial deponsidered, bortantly, thot afford th
vation itempresent agrly based oy considere
ted by Eurorty or sociais the propd with an eMarlier, 20forced inabthin the EU
pected expel holiday awn mortgagechicken, fisadequately
ngoing to remout clothes byer shoes; To sgether with frTo replace wo
ardsandtheM
112
epublic of ator for na small are
hboards anation indiceselect the isaggregatio
vationin
tion index utat (2002) ean countrprivation ratby most peohis indicatohe good or
s is specificreement wan prior sured as releva
ostat and pl exclusion”portion of “nforced ina13). Houseility to affoportfolio, d
nses way from hoe or rent, utsh or vegetay warm
move colour TVy new (not secpend a small ariends/family rn‐out furnitu
MaterialDepr
Moldova, cational repa index usin
nd multidims requires: final methoon; and rep
Europe
used to meconstructeies, and thete is defineople, to be dr considersservice, rat
c to time, pas found amrveys amonnt for the e
published as” indicator (“materially ability to affeholds are rd four or mdrawn from
ome ility bills, orarian equiva
V, washing mcond‐hand) onamount of mofor a drink/mure.
rivationIndic
countries hporting, anng a wider s
mensional pa process todology; theporting com
asure poveed an indexe index repoed as the endesirable ors the choicther than c
lace and pomong EU Mng the genentire popu
s a Europe (Eurostat, 2deprived” ford three oseen as “smore of theGuio et al.
r purchase ialent every
achine and tenes; Two paironey each wemeal at least m
ces
have the opd like the set of dimen
overty indito agree upe harmonismparable ind
rty and depx of non‐morted since nforced inar even necece that indchoose not
olicy contexMember Staeral populalation.
2020 sub‐in2015). The sindividuals,or more of tseverely mese items. T(2009), are
instalmentssecond day
elephone and s of properly fek on oneself
monthly; To ha
ption to United
nsions.
ces, the on a set sation of dices on
privation monetary 2010 by ability to essary to dividuals to have
xt and it tes. The ation to
ndicator standard , that is, the nine aterially The nine :
s) y
to add fitting f without ave
Chapter4PovertyDashboardsandtheMaterialDeprivationIndices
113
f) a washing machine g) a colour TV h) a telephone i) a personal car
330. All current items are addressing deprivation of the whole household as unit of observation. In order to aggregate the data, the nine items are combined at the individual level and then summed over individuals to form an aggregate index.34
331. The indicator is updated annually, based on the EU‐SILC instrument, and is publicly available on the Eurostat website. For some countries, the oldest data begin in 2004, while the most recent can be found for 2016. The indicator has been used by the European Commission, alongside their measure of monetary poverty and very low work intensity, to assess progress in reaching the EU‐2020 goal to “reduce the number of people at risk of poverty or social exclusion by 20 million by 2020 compared with 2008” (Eurostat, 2015).
332. A degree of overlap exists between those identified as materially deprived, income poor, and expenditure poor but there are also significant mismatches. In order to reach the EU‐2020 target, a focus on one aspect of poverty or social exclusion is not enough; rather, a multifaceted approach, backed with reliable data, is necessary. Data are also provided at a decomposed level, for such demographics as age, sex, household type, educational attainment, and country of birth. Figure 4.1 and
333. Figure 4.2 show some of these decompositions in graphical form. Figure 4.3 graphically highlights the differences between countries in 2014.
34 This is referred to as an “aggregated” indicator. This is distinct from “composite” indicators (such as UNDP’s Human Development Index), which first aggregates across people and then across these characteristics.
FigureSevere
FigureSeverecountr
Ch
4.1 ematerial
4.2 ematerialryofbirth,
hapter4Pove
deprivatio
deprivatio,EU‐28,20
ertyDashboa
onrate,by
onrateby013
ardsandtheM
114
ysexanda
household
MaterialDepr
agegroup,
dtype,edu
rivationIndic
EU28,20
ucationala
ces
10and20
attainmen
13
nt,and
FigureMapof
4.4.3
334. Whas beeof 13 itehave alpropose
335. Pe“enforc
f) Tog) Twh) To
ani) Toj) To
336. H
a) Tob) A c) Tod) O
35Availabdeprived
Ch
4.3 fseverely
Alternat
With a view en proposedems, six of so created ed list for m
ersonal iteced lack” tha
o replace wwo pairs of o spend a snyone o have reguo get togeth
ousehold it
o replace wmeal with o face unexOne week an
ble from http:/‐people [03/0
hapter4Pove
materially
tivemate
to revisingd by Guio, Gwhich are ia separate
material dep
ms: The peat does not
worn‐out cloproperly fitsmall amou
ular leisure aher with frie
tems: The h
worn‐out furmeat, chickxpected expnnual holida
//ec.europa.e02/2016]
ertyDashboa
ydeprived
erialdepr
g the officiaGordon, anncluded in e indicator fprivation ind
erson cannt simply refl
othes by somtting shoes,unt of mone
activities ends/family
household c
rniture ken, fish or vpenses ay away fro
eu/eurostat/w
ardsandtheM
115
dcountries
rivationr
al material dnd Marlier (the existingfor childrendicators for
ot afford bect a choice
me new (no, including aey each we
y for a drink
cannot affor
vegetarian
m home
web/gdp‐and‐b
MaterialDepr
s,2014,pe
rate
deprivation2012). Theig indicator, n (1‐15 yeathe whole
but would le):
ot second‐haa pair of all‐eek on ones
k/meal at le
rd:
equivalent
beyond/qualit
rivationIndic
ercentofp
rate, an alir suggestedseven of w
ars of age) wpopulation
ike to have
and) ones ‐weather shself without
east monthly
every secon
ty‐of‐life/seve
ces
population
lternative ind indicator which are newith 18 iteis as follow
e (i.e., a la
hoes t having to
y
nd day
erely‐material
n35
ndicator consists ew. They ms. The
ws:
ck is an
consult
lly‐
Chapter4PovertyDashboardsandtheMaterialDeprivationIndices
116
e) To avoid arrears (mortgage or rent, utility bills or purchase instalments) f) A computer and an internet connection (enforced lack: cannot afford but would like to
have) g) To keep the home adequately warm (enforced lack) h) A car/van for private use (enforced lack)
337. As an illustration, if we set the threshold at 5+ missing items (out of 13), the proportion of materially deprived people in the EU as a whole (EU‐27 weighted average) was 17.7 per cent in 2009, a percentage that is close to the current EU indicator of “standard” material deprivation (3+ items are lacking out of nine) of 17.1 per cent. A threshold of 7+ missing items (out of 13) leads to a material deprivation rate for the EU as a whole that is slightly higher than the current EU indicator of “severe” material deprivation (4+ lacked items out of nine): 9.2 per cent as opposed to 8.1 per cent.
338. The proposed multiple deprivations indicator for children is as follows:
Children’s items: The household cannot afford for at least one child to have (enforced lack):
Some new (not second‐hand) clothes
Two pairs of properly fitting shoes, including a pair of all‐weather shoes
Fresh fruits and vegetables daily
One meal with meat, chicken, fish or vegetarian equivalent daily
Books at home suitable for the children’s age
Outdoor leisure equipment
Indoor games
A suitable place to do homework
Regular leisure activities (sports, youth organisations, etc.)
Celebrations on special occasions
To invite friends round to play and eat from time to time
To participate in school trips and school events that cost money
One week annual holiday away from home
Household items: The household cannot afford:
To replace worn‐out furniture
To avoid arrears (mortgage or rent, utility bills, or purchase instalments)
A computer and an internet connection (enforced lack: cannot afford but would like to have)
To keep the home adequately warm (enforced lack)
A car/van for private use (enforced lack)
339. Box 4.4 represents the example of measuring material deprivation in the Republic of Moldova.
Chapter4PovertyDashboardsandtheMaterialDeprivationIndices
117
Box4.4 MaterialdeprivationintheRepublicofMoldova
In an effort to provide a broader picture of national poverty, the Republic of Moldova reports on three poverty measures: the proportion of the population at risk of income poverty, the proportion of households with very low work intensity, and a measure of material deprivation.
While it uses the EU‐SILC methodology, the Republic of Moldova’s indicators differ from those of EU‐SILC. Both measures focus on whether households can afford to avoid arrears, keep the house adequately warm, face unexpected expenses, eat protein (if desired), go on holiday for a week, and have a personal car, washing machine, television, and telephone. However, the Republic of Moldova considers those who answer “I do not know” to questions of affordability (2‐5) as not having financial difficulties in that indicator. Also, if households do not own any of the items (6‐9), it is assumed that they cannot financially afford it, as there is no additional information on affordability to buy these items. These assumptions are not made in the EU‐SILC measure.
Table 1 The items that make the object of the material deprivation in the EU and the Republic of Moldova
European Union Republic of Moldova Answer options
Your household could afford financially the follow:
In the last 12 months, your household had arrears conditioned by financial difficulties:
Avoiding arrears (in mortgage or rent, utility bills, or hire purchase instalments)
1. For paying utility bills 2. For bank credit reimbursement
1. Yes, once 2. Yes, several times 3. No
Your household could afford financially the follow:
Keeping the home adequately warm To keep the house adequately warm 1. Yes 2. No 3. I do not know
Face unexpected expenses Face unexpected expenses of 5000 lei 1. Yes 2. No 3. I do not know
Eat meat, fish or a protein equivalent every second day
To include in the diet meat or fish every second day (if desired)
1. Yes 2. No 3. I do not know
A week of holiday away from home once a year
A week of holiday away from home once a year
1. Yes 2. No 3. I do not know
Your household could afford financially the follow:
Quantity at the moment of the survey:
A personal car * Car, personal minivan Number of items ___
A washing machine * Automatic Washing Machine * Mechanical Washing machine
Number of items ___
A colour TV * TV Number of items ___
A telephone ** Telephone Mobile phone
Number of items ___
Source: Household Budget Survey (NBS, 2014) Notes: * The information is taken from Chapter 7 of the Main questionnaire “Durable goods in the
household” and the lack of these items does not represent that in reality the household has financial difficulties to afford them. ** The information is taken from Chapter 1 of the Main questionnaire “Household Dwelling”.
The Rehousehdeprivethose uover haand urb
Figure 1Share o
4.4.4
340. Ineconomof the “creationNorther
341. Th2004, 2measurrelative
342. Thacross lthat coveach smaveragethat wofunding
Ch
epublic of Mold cannot ed (i.e. theyused in the alf of the Reban areas.
1 of materially
Casestud
n the Unitemic circumst“social excln of an indern Ireland, S
he Governm2007, 2010,re of multipe scale: an a
he index ulocal areas ver a rangemall area ine score. Theould benefg streams.
hapter4Pove
Moldova reafford at le
y cannot affEU‐SILC. Es
epublic of M
y deprived
dy:Engli
ed Kingdomtances to guusion” concex of multipScotland, an
ment initiall and 2015.ple deprivaarea is chara
uses censusfor targetin of econom England, ae indices areit from spe
ertyDashboa
eports the east three oford at leasstimates ba
Moldova’s ho
people are
ishindex
m, interest uide urban cept in theple deprivand Wales. H
ly develope The Englisations at thacterised as
s and adming the provimic, social, aallowing thee then usedecial state
ardsandtheM
118
share of pof the ninest four of thased on theouseholds a
ea of residen
xofmulti
in measurpolicy resou EU in the tion (IMD) Here we foc
ed an index sh index of he small ars deprived if
inistrative dision of govand housingem to be rad to analyseinitiatives,
MaterialDepr
people who items) andhese items)e 2014 houare severely
nce, sex an
pledepr
ring geograurce allocat1980s. Thisthat has dius on the E
of deprivatdeprivationrea level. Df it is depriv
data to mavernment seg issues intoanked in ord patterns ofand to det
rivationIndic
o are mated those who. These defsehold budy deprived,
d househol
rivation
aphic variation came ws interest lefferent defnglish index
tion in 2000n is the GoDeprivation ved relative
ap differenervices. It co a single deder of deprf deprivatiotermine eli
ces
rially deprio are severfinitions aredget surveywith differe
ld size, 2014
tion in socwith the emed graduallyinitions in Ex.
0, and adjusovernment’sis measure to other ar
nces in depombines ineprivation srivation aroon, to identigibility for
ived (i.e., trely materiae the same y indicate thences by ru
4
cial and ergence y to the England,
sted it in s official ed on a reas.
privation dicators score for ound the ify areas specific
the ally as hat ural
343. ThdeprivadeprivaThe IMdeprivatrainingdomaincombinemploybarriersweightsof reseawere se
344. Insumma
Figure 4Overvie
Ch
he model oation, whication index,D brings toation takes g, barriers tn are combied to creatyment (22.5s to housins were selearch on preelected acco
Be “dommeasure
Measuresmall num
Be up‐to‐
Be updat
Be statist
Be availa
n order to crised in Figu
4.4 ew of the E
Source: Sm
hapter4Pove
of multiple h can be the IMD isogether 37 place: incoo housing aned to prodte the overa5%), health g and servcted basedevious indicording to te
main specifis of that for
major featmber of peo
‐date
table on a re
tically robus
ble for the
calculate thure 4.4.
nglish mult
mith et al. (20
ertyDashboa
deprivatiorecognizeds broader, idifferent inome, emploand servicesduce rankinall IMD 200and disabiices (9.3%) on theoretces and a coechnical crit
ic” and aprm of depriv
tures of depople or area
egular basis
st
whole of En
he indices o
tiple depriv
15a).
ardsandtheM
119
ons is basedd and meancluding hendicators thoyment, heas, living envngs for each07 after beiility (13.5%, living envtical considonsultation eria so that
propriate tvation)
privation (nas)
s
ngland at sm
of deprivati
ation meth
MaterialDepr
d on the idasured sepealth, educhat cover sealth and divironment, h domain. Tng weighte
%), educatiovironment (derations anprocess. Tht all indicato
to the pur
not just con
mall area le
ion seven s
hodology
rivationIndic
dea of distparately. Unation, and even dimensability, edand crime. he differened as follown, skills an(9.3%), andnd academihe indicatoors would:
rpose (i.e.,
nditions exp
vel in consi
steps are fo
ces
inct dimennlike the mcrime for ensions alonucation, skIndicators ft domains a
ws: income d training crime (9.3c work, thers for each
the best
perienced b
stent form.
ollowed, wh
sions of material example. g which kills, and for each are then (22.5%), (13.5%), 3%). The e results domain
possible
y a very
hich are
345. Si
346. NdescribpopulataccuratFigure domaindeprivafor each
FigureTheIn
36 Super Oon the N
Ch
ix local distr
1. Avethe
2. Avethe
3. ExteSOA
4. Locdistpop
5. Inco6. Emp
o single suming or comtion densitytely mapped4.5. Data an level, for dation withinh indicator
4.5 dexofMul
Source: Smi
Output Areas eighbourhood
hapter4Pove
rict level su
erage scoreSuper Outperage rankSOAs in a dent: the proAs in the coucal concenttrict’s mostpulation. ome scale: tployment s
mmary meaparing Engy and compd to revealare availabdecile, rank, that area rwithin each
ltipleDepr
th et al. (2015
are a geograpd Statistics sit
ertyDashboa
mmary mea
: the populput Areas (S: the populdistrict. oportion ofuntry. tration: thet deprived
the numbercale: the nu
asure is favoland’s 354 position. Ho the most le at the s, and score relative to th domain, w
rivation(I
5b). phy for the coe and across N
ardsandtheM
120
asures of th
ation‐weighSOAs)36 in alation‐weig
f a district’s
e populatioSOAs that
r of people umber of pe
oured over local autho
owever, onc(and least) small area statistics. The other arwith specific
IMD),2015
ollection and pNational Stati
MaterialDepr
he IMD have
hted averag district. hted averag
s populatio
on‐weightedt contain e
who are inceople who a
another asority districce the indexdeprived alevel, at bThe former treas, while tc numerator
5
publication ofstics (ONS).
rivationIndic
e been prod
ge of the co
ge of the c
on living in
d average exactly 10%
come depriare employ
s there is nots because x is calculatareas in Engoth IMD letwo statistithe latter strs and deno
f small area st
ces
duced:
ombined sc
combined ra
the most d
of the ran% of the d
ved. ment depriv
o single besof the diveed the datagland, as shevel and incs relay dattatistic is caominators.
atistics. They
cores for
anks for
deprived
nks of a district’s
ved.
t way of ersity in a can be hown in ndividual ta about lculated
are used
4.4.4.1
347. Wshould
4.4.5
4.4.5.1
348. Theconomthe exte
349. Uimprovemay alsomitted
350. Min termerrors.
351. Thacross cbetwee
4.4.5.2
352. Threflect ojoblessnthat the
Ch
1 Possib
While the IMbe consider
The publscores isdone on
There arecompararelative n
The data(howevethe data
While ththe “rawrobustne
Because and Wale
Assessm
1 Advan
he clear admic deprivatent to whic
sed togethe the accurso reflect wd from mon
Material deps of survey
he materiacountries. Cen currencie
2 Disad
he materiaother relevaness shows e United Kin
hapter4Pove
blelimita
MD is a usefured:
lished ranks not straigan absolute
e issues coble for onenature of th
for each or, most for will always
e decile, raw” empiricaless.
separate ines, compari
mentofm
ntages
dvantage oftion that moh poverty s
her with incracy of povewealth or penetary meas
privation indy length and
l deprivatioComparisones.
dvantages
l deprivatioant dimensyet a differngdom Indic
ertyDashboa
ations
ul tool for m
s and decilhtforward. e scale.
ncerning dye year, reahe index.
of the indicathe 2015 lag somew
ank and scol data is lim
ndices are sons betwe
ateriald
f material donetary povtatistics ma
come poveerty assessmermanent insures.
dices have fd complexit
on index, ons do not r
s
on index is ions of povrent patternces of Mult
ardsandtheM
121
measuring d
es are pureNo statisti
ynamic comal comparis
ators are nIMD are fro
what behind
ore statisticsmited, whic
produced feen the four
eprivatio
deprivationverty measatch experie
erty measurments. For ncome, inso
fewer data ty, and may
nce validatequire pric
a limited perty, nor thn of deprivaiple Depriva
MaterialDepr
deprivation
ely relativec is cardina
mparisons: wsons over t
ot from a som the 201the current
s are readilh complicat
for Englandr are unfeas
onindice
indices is ures overloences of pov
res, materiaexample, m
ofar as these
requiremeny have lowe
ed, is easy es, inflation
proxy for poheir joint disation, as miation do inc
rivationIndic
in England,
, so the intal, so comp
while the intime are co
single consi12/13 tax yt situation.
ly availabletes assessm
d, Northernsible.
es
that they ook. They caverty.
al deprivatmaterial depe are releva
nts (than mer non‐samp
to computn, or purch
overty. By dstribution. Fight health clude other
ces
, several lim
terpretationparisons ca
ndex for anomplicated
stent pointyear). Furth
, the availament of the
n Ireland, S
capture aspan thereby i
ion measurprivation mant to the p
monetary mepling measu
te and to casing powe
design, it dFor exampleor educatiodimensions
mitations
n of the nnot be
n area is by the
t in time ermore,
ability of e index’s
cotland,
pects of improve
res may measures poor but
easures) urement
compare er parity
does not e, quasi‐on. Note s.
Chapter4PovertyDashboardsandtheMaterialDeprivationIndices
122
353. For this reason, while a material deprivation index may be a very welcome component of a dashboard or an index, it cannot serve as a stand‐alone measure of poverty in all its dimensions.
354. Some statistically valid elements of the material deprivation index may be problematic in terms of policy prescriptions—car ownership being one such example. Thus, the methodology needs to incorporate consultative input and policy considerations into indicator design, while ensuring that the final measure satisfies statistical criteria.
4.5 Conclusion
This chapter describes the processes and principles of establishing comparable indicators of poverty and social exclusion and measures of material deprivation. The requirements for comparable dashboards are briefly described, including the advantages and disadvantages of such dashboards. As in the case of comparable dashboards, the development of official material deprivation indices also requires certain processes and principles to be followed. These indices can complement monetary poverty measures, and are recommended as one of the comparable indicators on a poverty dashboard. A number of case studies demonstrate the measurement of deprivation, as users may find useful the experiences of some countries in tailoring deprivation measures to their data and contexts.
Chapter5MultidimensionalPovertyIndices
123
5 MultidimensionalPovertyIndices
5.1 Introduction
355. Some national and regional policies already address non‐monetary deprivations such as housing, health, education, and services. But few look at the overlap across non‐monetary deprivations, such as the EU material deprivation measure. The new multidimensional poverty measures do both. Multidimensional poverty measures are being adopted by many countries including Armenia, Chile, Colombia, and Mexico as official permanent statistics that complement monetary poverty measures. When comparable multidimensional poverty indices (MPIs) are available across countries, for example in the MPI published by the Oxford Poverty and Human Development Initiative (OPHI) and UNDP, we can compare national experiences. This chapter explains how to construct multidimensional poverty measures and what their interpretation provides for policy. It also gives many examples.
356. The chapter presents MPIs as a useful and popular complement to countries’ national monetary poverty measures, as can be seen from the examples of countries that use national MPIs (and their associated dashboards) as official statistics. Step by step guidelines are provided on how to design and use a national MPI for policy. The chapter also introduces some internationally comparable MPIs, for example the one published by UNDP, and shows what value‐added emerges from comparable MPIs. By introducing these measurement techniques, the chapter provides a set of resources on how non‐monetary aspects of poverty can be measured and monitored. It also shows the value‐added and challenges of different measurement approaches.
357. Naturally, one cannot discuss measurement without discussing data. So, the chapter also addresses the data needs. A core data question concerns whether UNECE countries will develop a comparable regional MPI. At present, it is not possible to do so using existing data. The chapter therefore proposes that each country develop a national MPI that suits its national data sources and policy objectives. However, importantly, it proposes that they do so with an aim towards an eventual harmonization using a core subset of dimensions and indicators. These could be used to make a UNECE MPI and dashboard. Considering recent examples and participatory consultations, a core subset of dimensions is likely to include living standards, services, health, education, and the lived environment.
Recommendation 23: Each country should develop a national MPI that suits its national data sources and policy objectives. It is desirable that the national MPI includes the dimensions of living standards, services, health, education, work and the lived environment.
358. If national priorities are met (these have priority), and if in time, harmonized data are available, then for low marginal costs, both national and comparative measures can be built for UNECE countries. Comparing country experiences seems to be of considerable interest to UNECE member states. With both national and regional aims in mind, this chapter aims to support countries’ exploration of rigorous multidimensional poverty measures, and to encourage the creation of data sources that would permit the generation of a regional MPI based on comparable indicators of non‐monetary poverty. With both national and regional
Chapter5MultidimensionalPovertyIndices
124
aims in mind, this Guide aims to support countries’ exploration of rigorous multidimensional poverty measures, and to encourage the creation of data sources that would permit the generation of a regional MPI based on comparable indicators of non‐monetary poverty.
359. In the case of monetary poverty, national income poverty measures are used for national poverty reduction policies, while cross‐national studies are conducted that draw on comparable measures such as the PPP$3.10/day poverty measure to elucidate good practices that would be relevant to other UNECE countries. A similar structure is proposed for multidimensional poverty, with national measures providing the basic tool for national policymaking, and a restricted yet comparable multidimensional measure providing insights and lessons learned across national boundaries.
360. Some measurement considerations are similar to monetary poverty: in particular, Recommendations 1‐4 of Chapter 3 also pertain to multidimensional poverty. These refer to the unit of identification, the unit of analysis, the need to include overlooked populations in household survey sampling frameworks, and disaggregation by key population subgroups.
5.2 Overview
361. Multidimensional poverty indices are being developed by many countries as official national poverty statistics. Armenia, Bhutan, Colombia, Chile, Costa Rica, the Dominican Republic, El Salvador, Ecuador, Honduras, Mozambique, Pakistan and Panama among others have official multidimensional poverty indices (MPIs) that complement their official monetary poverty statistics and are updated and reported regularly alongside monetary poverty measures (all country documents are linked from www.mppn.org). Mexico has a single official poverty measure, which became multidimensional in 2009, and includes income and six non‐income components. Countries such as Tunisia and Turkey are designing national MPIs. Academic studies in the United States, Germany and elsewhere are exploring these measures, and UNDP has published studies of social exclusion that implement MPIs in the analysis (Bartels and Stockhausen 2017, Brucker et al 2015, Nowak and Scheicher 2017, Suppa 2017, UNDP 2011, Wagle 2014).
362. This chapter introduces MPIs and shows how they add value to a monetary poverty measure or a dashboard. Differently from the index of material deprivation discussed before, MPIs reflect a view of poverty that remains multidimensional, being grounded in Sen’s capability approach. In contrast, measures of material deprivation focus on a single underlying phenomenon, material deprivation, and seek to describe it using various indicators. Because of this, different techniques are used to build the index of material deprivation. In practice, material deprivation can be one indicator within an MPI.
363. The methodology underlying the MPI is based on Alkire and Foster (2011) and offers a high degree of flexibility in the choice of indicators. These indicators can be tailored to suit the specific requirements of each country and reflect the priorities of policymakers. The MPI can be used for a multitude of policy purposes including: targeting of social and economic policies, monitoring their impact and implementation, coordination among different decision makers, assessment of sub‐national differences in development, graduation of social protection schemes, and informing socially responsible investments.
Chapter5MultidimensionalPovertyIndices
125
364. The MPI can be a particularly useful tool to assess how countries meet the SDGs. The first goal of the SDGs is to eradicate poverty in all its dimensions, meaning that the SDGs focus on multidimensional poverty. MPIs based on the Alkire‐Foster method are being reported by many countries for Indicator 1.2.2 (See Section 5.9).
365. There are two kinds of MPIs. National MPIs are like national monetary poverty measures. They reflect national priorities and are constructed using national datasets. But they cannot be compared. A regional or global MPI, like the global income poverty measure of PPP$1.90/day, is comparable across countries. For example, the OPHI/UNDP Global MPI covers a similar number of developing countries as the World Bank’s PPP$1.90/day measure, drawing on national and international datasets, the Economic Community of Latin America and the Caribbean (ECLAC) published a regional MPI in their 2014 Social Panorama, and Alkire and Apablaza (2016) created a preliminary MPI across European Countries using EU‐SILC data (See Box 5.3). MPIs in either comparable or national forms can be used by countries to report on SDG target 1.2.
Recommendation 24: UNECE countries should report a Multidimensional Poverty Index against SDG target 1.2. In the short term, countries can report existing national MPIs or the value of their global MPI published by UNDP.
366. In this sense, the flexibility of the Alkire‐Foster method used to build any MPI allows the index to capture national and international concerns of poverty and development (e.g., national development plans and SDGs). The current MPIs that are in place include a range of indicators pertinent to SDG goals such as health, education, living standards, social inclusion, violence, and employment, among others.
367. This chapter describes methodological issues in building an MPI, starting with the steps needed to set the unit of identification, dimensions and indicators with their respective deprivation cut‐offs and weights, and poverty cut‐offs. It also illustrates the advantages and disadvantages of some methodological issues in building an MPI.
5.3 Requirements
368. Building an MPI may seem more complicated than an income poverty measure because it is new. So it is useful to clarify three common misconceptions from the start. The first is that a MPI is more data hungry than a monetary poverty measure because it covers more dimensions. In fact, whereas around 500 items from the same survey are often used to build a consumption aggregate, an MPI is ordinarily built from between 30 and 50 survey questions that have to come from one and the same survey. Thus the Atkinson Commission Report “Monitoring Global Poverty” (World Bank, 2017) rightly observes that the MPI is less data intensive than monetary measures: “The creation of the overlapping poverty measure, or of the more general measures of multidimensionality developed by Alkire and Foster (2011a), in one sense raises the stakes with regard to data requirements. In order to ascertain the extent of overlap of deprivation across dimensions, it is necessary to have a data source at the level of the individual or household covering all relevant dimensions. At the same time, the number of questions required per dimension may be considerably less in the case of nonmonetary indicators than is the case with the measurement of consumption for the monetary policy indicator. The multidimensional poverty indicator for Colombia is
based oRica onmuch mconsumassumepovertysignifica
369. A MPI refinflationcountrydatasetonline tnationaevery d
370. A ministrilegitimamay ocstatisticcoordinrelevanpolicy apowerftogetheon.37
5.4 S
5.4.1
371. It whethewill be the unipercentcan be interpreage coh
37 The manational
on some 38 77 survey more extenmption recaed that a ny and emploantly faster
second legflects povern, nor impuy can releast. It can alsthe computal MPIs—soetail.
third queries and seate proceduccur at the cs users (ponating bodint insights tactions to ul tool forer in budge
Stepsto
Prelimin
is first necer each persreported. Lt of identiftage of peothe unit ofeted, howehorts.
agazine DimenMPIs are bein
Ch
survey quequestions. nsive. For ell list of somon‐monetaoyment que.
gitimate qurty directly, utation for se its updateso be donetational algo students, t
ry is how aectors. Likeure, clarificanational or
olicy makerses) in meathe measurreduce povr governanct allocation
buildan
narystep
essary to dson is poorLike in the cfication becople who arf identificatver compar
nsions publishng used in pol
apter5Multi
estions, thaThe informexample, tme 450 itery approacestions are
ery is whetit does notrent. In pred official ne transpareorithms (STthink tanks
n MPI can other indation of dar regional ls from diffeasurement re can provverty. Expece, becausn, targeting,
nMPI
p
etermine thr or not, ancase of moncause of dare poor. If dtion, in whirable data
hes country caicy. See http:/
idimensional
126
t for Pakistmation requhe 1993–9ms (Gibsonch is more time consu
ther the MPt require adractical termnational MPently. MexicTATA, ADePs, and analy
inform poldicators, thta sources, evel. It hasrent relevadesign. Thivide, such terience to de it addre, policy des
he unit of idnd the unit netary poveta constraidata are avach case genare rarely a
ase studies an//www.mppn
PovertyIndi
an on 54 suired to calc4 survey fn 2005, 137data‐demauming to co
PI takes londjustment (ms, after thPI two weekco, EcuadoPT, R, SAS oysts can cle
licy, becaushe developand a clea
s proven trent sectors ais facilitatehat these adate has shsses interlisign and coo
dentificatioof analysis rty, most conts, and moailable at thnder disaggavailable at
nd synthesizes.org/dimensio
ices
urvey questculate consufor Cambod7). Therefornding”. Anollect, other
nger to comsuch as for he MPI is iks after it rer, Colombiaor SPSS) neearly unders
se is implicment of ar reportingemendousland levels os their undare translathown that inked dimeordination,
on that will by which tountries usost aggregahe individuagregation ca the individ
s other documons/editions/
tions, and foumption is tdia had a e, it shouldnd while mr indicators
mpute. Becaurban‐ruranitially deseceives the a, and otheeded to mastand and r
cations affean MPI reqg frameworky useful toof governmederstandingted effectivthe MPI caensions of evaluation
be used to the povertye the houseate in termal level, thean be meandual level a
ments that sho
or Costa typically detailed d not be monetary may be
ause the al prices, igned, a cleaned ers post ake their replicate
ct more quires a k. These involve ent, and g of the vely into an be a poverty , and so
identify y figures ehold as s of the e person ningfully cross all
ow how
372. O(2015) d
5.4.2
a)
b)
c)
d) e)
f)
5.4.3
a)
b)
c)
373. Fithe bas
Once this hdescribe the
Identific
Defining thmeasure. identificatioSetting theconsideredApplying thindicator. Selecting thCreating thor her deprDeterminindeprivationmultidimennot accordi
Aggregat
Computingpoor in thealso called Computingdeprived. Tby the totpoverty, alsComputingindices: H tcensored dpoor peopl
igure 5.1 beic example
Ch
as been ace following
cationoft
he set of iData for aon. e deprivatio sufficient (he cut‐offs
he relative whe weightedrivation scong (normatins a persnsionally poing to the se
tion
the propore populationthe inciden the averagThis entails tal numberso sometim the M0 mtimes A. Aneprivation e experienc
elow showsof computi
apter5Multi
ccomplishedsteps for th
thepoor
ndicators wall indicato
n cut‐offs f(normativelto ascertai
weight or vad sum of dere. vely) the pson needsoor, and ideelected pov
rtion of peon. This is thnce of multidge share (Aadding up tr of poor mes called thmeasure (ornalogously, scores, whice, divided
s the differeng a MPI in
idimensional
127
d, Alkire, Fhe construc
r
which will ors need t
for each indy) in order in whether
alue of eacheprivations
overty cut‐s to expeentifying eaverty cut‐of
ople who hahe headcoudimensionaA) of weighthe deprivatpeople. Thhe breadth or MPI) as tM0 can bech is also thby the tota
ent steps nn practice.
PovertyIndi
Foster, Sethction of the
be consideto be avai
dicator, namto be non‐deach perso
h indicator,for each pe
‐off, namelyerience in ch person aff.
ave been idnt ratio (H)al poverty. ted indicattion scores his is the iof poverty. he producte obtained he sum of tl population
eeded to b
ices
h, Santos, MPI:
ered in thelable for t
mely the levdeprived in on is depriv
such that terson, whic
y, the proporder to
as multidim
entified as ) of multidim
ors in whicof the poorntensity of
t of the twas the meahe weighten.
uild a MPI.
Roche, and
e multidimethe same
vel of achieeach indicaved or not
these sum tch can be ca
ortion of wo be conmensionally
multidimenmensional p
ch poor peor and dividinf multidime
wo previousan of the vd deprivatio
Box5.1 ill
d Ballon
ensional unit of
evement ator. in each
o one. alled his
weighted nsidered poor or
nsionally poverty,
ople are ng them ensional
s partial ector of ons that
ustrates
FigureStepsf
Step 0
Step 1
Step 2
Step 3
Step 4
Step 5
Step 6
Step 7
Step 8
Step 9
5.1 formeasur
0
•Determ•e.g., Ho
1
•Define •e.g., ye
2
•Set Dep•e.g., decohort.
3
•Apply t•Find ou
4
•Select t•each inone (10
5
•Create •Add up
6
•Determ•An indiv
7
•Calculat•This is t
8
•Calculat•This is cdividing
9
•Calculat•This is tStep 8
Ch
ringaMult
mine unit of idousehold or I
a set of indicars of schoo
privation Cutprived if yea
he Deprivatiut whether o
the relative wdicator will h00%).
the weightedthe weights
mine Poverty vidual is iden
te the heacothe proportio
te the intenscalculated byg by total num
te the MPI (othe product o
apter5Multi
tidimensio
dentificationndividual
cators ling, housing
t‐off ars of schooli
on Cut‐off r not each pe
weight of eachave a relativ
d deprivatios of all of the
cut‐off (x%)ntified as mu
ount ratioon of the pop
sity of povery adding up tmber of poo
or M0)of the headc
idimensional
128
onalPover
n
g, decent wo
ing less than
erson/house
ch indicatorve weight, a
n score (rang indicators in
ultidimension
pulation who
ty among ththe deprivatir people
count ratio o
PovertyIndi
rtyIndex
ork, health st
national com
ehold is depr
nd these we
ging from 0 tn which peop
nally poor if
o are poor
e poorion scores of
f Step 7 and
ices
atus
mpulsory lev
rived in each
ights genera
to 100%). ple are depr
the deprivat
f all poor per
the intensity
vel for each a
indicator
ally sum up to
ived.
tion score > x
rsons and
y of poverty
age
o
x%
of
Chapter5MultidimensionalPovertyIndices
129
Box5.1 AnexampleofhowtocomputetheMPI
Suppose we have a society of three individuals: Anna, Diana, and Victor. The MPI we seek to construct has the following indicators nested within three dimensions
Dimensions of poverty Indicator Deprived if… Weight
D1 a … 1/6
b … 1/6
D2 c … 1/6
d … 1/6
D3
e … 1/18
f … 1/18
g … 1/18
h … 1/18
i … 1/18
j … 1/18
Data are collected for these three individuals, and a deprivation matrix is constructed as show in the table below (indicator weights are shown in parenthesis). Values of 1 in the table indicate deprivation in this component.
Indicators Deprivation
Score a (1/6)
b (1/6)
c (1/6)
d (1/6)
e (1/18)
f (1/18)
g (1/18)
H h(1/18)
i (1/18)
j (1/18)
Anna 1 1 0 0 1 0 0 1 0 0 4/9 = 0.44
Diana 0 0 1 0 1 0 1 1 0 0 1/3 = 0.33
Victor 0 1 0 0 0 0 0 0 0 0 1/6 = 0.17
If the poverty cut‐off is, for instance, 33%, then both Anna and Diana are identified as multidimensionally poor because their deprivation scores are 33% or higher, while Victor is not.
Assuming equal sampling weights (this is often not the case in sample data), we get the following measures:
1. The headcount ratio (H) is 2/3 = 0.66. That is, 66% of the population is MPI poor.
2. The intensity of poverty among the poor (A) is obtained as the average deprivations among the poor, or (0.33+0.44)/2 = 2/5 = 0.39. That is, the poor are, on average, deprived in 39% of the indicators.
3. The adjusted headcount ratio M0 (or MPI) is H x A = 0.66 x 0.39 = 0.250.
5.5 K
5.5.1
374. Tothe depwithin apoor orpovertyBut in t
375. Inthe houexercisereveal a
376. Wdecompassess ithat arechildren
377. Wthat onbe inclhousehdeprivamembe
5.5.2
378. DweightiDimenspolicymthe Glohealth,
379. Soinclude
38 Each o39 For a s40 More iNations E41 See Alkcomposit
Keydeci
Unitofid
o build an Mprivations oa householdr non‐poory index. Usuhis case sex
n multidimeusehold. Boes, and eacat least som
When the upose by genintra‐housee relevant fn and for th
When the une householuded,41 andold membeation may afers can read
Dimensi
imensions ng structusions for a nmakers, civilobal MPI (seducation,
ome examp:
f these decisioystematic trenformation oEconomic Comkire and Santotions.
Ch
isions38
dentifica
MPI, a preliof an indivd or other ur. The sameually for mox disaggrega
ensional meoth optionh has adva
me of the int
unit of idender, age, oehold patterfrom cradlehose who ar
nit of identld membersd allows cer is not hffect their od essential m
ons
are concepre and to national MPl society, resee Sectionand living s
ples of dim
ons is elaboraatment see An the importammission for Eos (2014) for a
apter5Multi
tion
minary stepvidual, or wunit, to idene choice is onetary povation, for ex
easures, thes have beentages and tra‐househo
entification occupation,rns of povee to grave, sre above a c
ification is ts’ deprivatiompensatioighly educaown povertmaterials.
ptual categexplain th
PI should beesearchers n 5.6) are standards.
mensions us
ated in Alkire elkire et al. (20ance of statistEurope (2015ban example of
idimensional
130
p is to choowhether to ntify all memmade in certy, housexample, can
e unit of iden used indisadvantaold inequali
is the ind, and otherrty. Howeveso separatecertain age.
the househon has on oon in someated but yoy condition
gories that he measure selected inand statistigrouped in
sed in offic
et al. (2015). 015), Chapter tics and indicab). f analysis of th
PovertyIndi
ose whethecombine inmbers of a constructingehold incomnnot be mea
entification official naages.39 In boities.40
dividual, it r characterier, it may be measures
hold, this acothers, it ae indicatorsounger one much less,
can be ure to the n consultaticians. For nto three e
cial multidi
7. ator on gende
he global MPI
ices
r to identifyndividual lehousehold g an incom
me or consumaningfully p
n is normallational MPIoth cases th
is possiblistics. That be difficult tmay be mo
cknowledgellows somes. For instaes are, the, because th
sed to facpublic andion with staexample, thequally wei
imensional
r equality is a
for househol
y poverty bevel characequally as a
me or consumption is asperformed.
ly the indivIs and in rhere is pote
e to meanis, it is posto define inore approp
s the sharee child indicance, if an elder eduheir younge
ilitate a pad to policyakeholders,he 10 indicghted dime
poverty m
vailable in the
ds with differ
based on cteristics all being umption ssessed.
vidual or research ential to
ningfully ssible to dicators riate for
ed effect cators to elderly cational er family
articular ymakers. such as cators of ensions:
measures
e United
ing child
Chapter5MultidimensionalPovertyIndices
131
1. Health 2. Education 3. Work 4. Housing 5. Living standards 6. Basic services 7. Lived environment 8. Social security 9. Personal security (safety from violence) 10. Food security 11. Childhood and youth
380. The choice of dimensions in Different National MPIs is presented in Box 5.2.
Box5.2 ThechoiceofdimensionsindifferentnationalMPIs
As mentioned previously, the Alkire‐Foster methodology for MPI construction allows a large degree of flexibility in the choice of dimensions, the numbers of dimensions, indicators, and weights. Different countries have built national MPIs in keeping with their national development agendas, in order to target specific groups of the population and monitor social protection schemes. The table below shows the different dimensions chosen by national MPIs around the world.
Table 1 National MPI dimensions worldwide
Country Dimensions
Armenia (1) Education, (2) Health, (3) Labour, (4) Basic Needs, (5) Housing
Chile (1) Education, (2) Health, (3) Work and social security, (4) Basic standard of living
Costa Rica (1) Education, (2) Health, (3) Work and social security, (4) Basic standard of living
Colombia (1) Education, (2) Childhood and youth, (3) Work, (4) Health care, (5) Housing and public services
Dominican Republic
(1) Education and childcare, (2) Health, (3) Work and Livelihood (4) Housing and environment (5) Digital Divide and Social Relations
Ecuador (1) Education, (2) Health, water and nutrition, (3) Work and social security, (4) Housing and public services
El Salvador (1) Education and childhood, (2) Health and food security, (3) Work, (4) Housing, (5) Security and environment
Mexico (1) Education, (2) Access to health care, (3) Access to food, (4) Access to social security, (5) Housing, (6) Basic home services, (7) Income
Panama (1) Health, (2) Education, (3) Work, (4) Housing, Basic Services & Internet Access, (5) Environment and Sanitation
ECLAC (1) Housing, (2) Basic Services, (3) Living Standards, (4) Education, (5) Employment and Social Protection
Pakistan (1) Education, (2) Health, (3) Living Standards
5.5.3
381. Inidentifyshould 7‐20 instronge
382. NExceptilocation
383. Dinclude SDG indsuch asengagin
384. Dcommuan indicmust bpropertpeople and dat
385. Hovercroso on, househ
386. Inemploypensionhousehbe comindicatodeprivahouseh
387. Thexamplseconda
Indicato
ndicators ary whether abe limited, ndicators. Rest indicator
ormally thonally, admn, if the me
eprivation compulsordicator defis a constitung poor peo
eprivation unications ocator with be taken inties like thedo not affeta reasons (
ousehold owding, wasall draw uold level de
ndicators suyment, qualns, and so oold is the u
mbined to dor. For examation, or if aold types. E
If no mehouseho
If any sccompulsoschool ag
If all houdeprived
he design oe, definingary school
Ch
rsandde
re the variaa person is for ease of
Redundancyrs (Alkire et
he data foministrative rging mana
cut‐offs nory years of nitions. In ttion, nationople and com
cut‐offs areof trends ova relative cn explaining focus axiomect the MPfor example
level indiste disposapon houseeprivations
uch as mality of workon, are normunit of identdefine the mple, a houall persons Examples of
ember of tld is deprive
chool agedory level oged children
usehold me.
of such indg as deprivcreates a
apter5Multi
eprivatio
bles in the deprived (f communicy analysis at al, 2015, C
or all indicor satelliteges to retai
ormally draf schooling,the case of nal developmmunities.
e absolute, ver time, becut‐off (likeg changes ms (so chanPI). It may ae if variable
icators sual, electricityehold indicashould be a
lnutrition, k, social secmally availabtification, inhousehold
usehold migdo, or somf complex in
the househed.
d child is nf schoolingn are consid
embers are
icators shoved anyonesituation c
idimensional
132
oncut‐off
matrix thaor not) in tcation and and other Chapter 7).
cators is a data may bin a very hig
aw upon co or age‐spenational Mpment plan,
not relativecause the e income) isover time
nges in the aalso contribes are binar
ch as wy, material ators. If anattributed to
health, yeacurity benefble for certanformation d and all itght be deprime proportiondicators ar
hold has c
not attendig, the housdered as no
unemploye
ould ensuree in the hocreates eff
PovertyIndi
fs
t are used, that indicatpolicy focustechniques
available frbe merged gh proportio
ommonly aecific standPIs, they m, or the res
ve. This is tothresholds s included . It also mattainmentsbute to a sty or catego
ater, sanideprivation
n individual o all househ
ars of schofits, health ain or all indfrom relevats membersived if any pon. The indre:
ompleted
ng school sehold is dn‐deprived
ed or unde
e that they ousehold wectively a
ices
with a deptor. The nus. In practics can be u
rom the sausing the hon of the or
agreed standards for may also drawsult of parti
o ease the are fixed, ain the MPI,means that s of non‐potronger merical).
tation, hon, access tois the uni
hold memb
ooling or sinsurance, dividuals in ant househs as deprivperson withicator must
lower seco
and has neprived. Ho.
r‐employed
are responwho did nostock indic
privation cumber of ince, most MPused to se
ame data household iriginal hous
ndards. Themalnutritionw upon docicipatory pr
interpretatand are intu, considerathe MPI
oor or non‐dasure for t
ousing cono the internit of identier equally.
school attedisability ba householold membeved or not hin it expert be define
ondary scho
not compleouseholds
d, the hous
nsive to poot completcator, beca
ut‐off, to dicators PIs have lect the
source. d or GIS seholds.
ese may , or the cuments rocesses
tion and uitive. If ble care satisfies deprived echnical
nditions, net, and fication,
endance, benefits, ld. If the ers must in that iences a d for all
ool, the
eted the with no
ehold is
licy. For e lower use the
uneducstatus t
5.5.4
388. Aso as tplausibfor examin intrinfor polisuch thexceptidimensparticulperformweights
Recommmultidimfrom thpolicy r
389. OpresentThe issusuch asmoneta
cated will likthrough the
Weights
s indicated to be openle range ofmple, that ansic value. Ecy, to the ehat their weon of Chilions, and larities requmed and res (Alkire and
mendation mensional his can be nelevant com
Other inputsted in deveue of weighs imputing pary poverty
Ch
kely includeeir life cours
by Sen (19n to public f weights. NachievemenEqual weighextent that eights can be, have usequal w
uired the meported, to d Santos 20
25: It is repoverty indormatively mparisons a
s into the lopment plahting is no mprices for nomeasures.
apter5Multi
e many adulse, so it will
96), the wedebate; an
Normally, thnts in healthts across dAtkinson ebe roughly sed a nesteights acr
modificationensure th
014, Alkire e
ecommenddex, and ejustified. Rare robust t
final weighans – as wemore challeon‐market g
idimensional
133
lts who canonly chang
eighting strund further, he relative h, educatiodimensions et al. (2002)equal. In fated weightross indican of this struat the finaet al, 2015, C
ed to use qual weighobustness to a range o
hting structell as particinging for thgoods or ad
PovertyIndi
not changege when the
ucture shoukey compweights re
on, and livinalso ease t) recommenact, all offict structure,ators withiucture. Alsoal MPI is roChapter 8).
equal weights among tests shouldf plausible w
ure includeipatory exehe MPI thandjusting rur
ices
e their educey leave the
uld be expliparisons muflect normag standardshe interprend that dimcial statistic, with equin dimenso, robustneobust to a
ghts for thindicators d be reportweights.
e national prcises with n for other ral and urba
cational depe household
cit and tranust be robuative assesss are roughetation of thmensions bes to date, wual weightsions—unlesess tests arerange of p
e dimensiounless a ded, to see w
policies – opoor commpoverty indan poverty
privation d.
nsparent ust to a sment—hly equal he index e chosen with the s across ss data e always plausible
ons of a eviation whether
often as munities. dicators, lines for
Chapter5MultidimensionalPovertyIndices
134
Box5.3 MultidimensionalPovertyIndicatorsinEurope:EU‐SILC
Alkire and Apablaza (2016) calculate an MPI for Europe based on the Alkire‐Foster (AF) methodology drawing on existing Europe 2020 indicators, as well as on indicators of health, education and the living environment. The MPI (which consists of 12 indicators) can be compared across time and space (Table 1).
Table 1 Dimensions, indicators and weights applied for generating a MPI
Dimension Indicator Variable Weight
Income AROP 1/6
Employment Quasi‐Joblessness 1/6
Material Deprivation
Severe Material Deprivation
1/6
Education Completed Primary Education
1/6
Environment
Noise 1/24
Pollution 1/24
Crime 1/24
Housing 1/24
Health
Fair Health 1/24
Chronic Illness 1/24
Morbidity 1/24
Unmet Medical Needs
1/24
The results (shown in Table 2) suggest a decline in multidimensional poverty in Europe from 2006 to 2012. However, while the MPI headcount fell from 10.04% in 2006 to 8.81% in 2012, the intensity of poverty among the poor people remained largely unchanged.
Table 2 Multidimensional Poverty in Europe 2006‐2012, k=34% (Level and percentage of individuals in EU countries with consistent data – linearized std. errors in brackets)
2006 2007 2008 2009 2010 2011 2012
Multidimensional poverty (Mo)
0.0484 0.0443 0.0418 0.0413 0.0419 0.0424 0.0429
(0.0012) (0.0011) (0.0012) (0.0012) (0.0011) (0.0011) (0.0011)
Headcount ratio (H) 10.04% 9.24% 8.77% 8.63% 8.67% 8.75% 8.81%
(0.0012) (0.0012) (0.0013) (0.0013) (0.0013) (0.0012) (0.0013)
Intensit
Contrib
Income
Employ
Materia
Educati
Environ
Health
SoN
Switzer
With thwhich c
H
Ed
H
Other require
5.5.5
390. Cin whicare apppoor or
391. Acut‐off.the case
392. Pothe weiof factopoverty
ty of poverty
ution of each
yment
al deprivation
on
nment
ource: AlkirNote: Belgiurland nor inc
he update ican be used
ousing (ove
ducation (anational eimproveme
ealth (disab
possible ind for good
Poverty
ut‐offs for h a person plied to the r non‐poor.
s in the cas For exampe of the nat
overty linesighting struors, includiny assessme
Ch
48.18%
(0.0006)
dimension to
24.23%
18.40%
16.13%
17.94%
11.80%
11.50%
e and Apabum, Bulgarcluded.
in the EU‐S to build na
ercrowding,
assessmenteducation ent over the
bility, activit
ndicators inliving stand
cut‐off(s
multidimenor househodeprivation
se of incomple, an extretional MPIs
s are easiescture of indng normativnts, or obs
apter5Multi
47.99%
(0.0006)
total multidim
24.58%
18.68%
15.83%
17.46%
12.07%
11.38%
blaza (2016)ria, Ireland
SILC surveyational MPIs
, accommod
s of an indschemes,
e previous i
ty)
nclude inddards.
)
nsional poveold must ben scores, an
e poverty, eme povertyin Ecuador
st to presedicators andve assessmservation o
idimensional
135
47.73%
(0.0006)
mensional pov
25.23%
18.31%
15.56%
17.90%
11.34%
11.66%
based on Ed, Croatia,
s after 201s. These inc
dation, room
dividual’s ovocationa
ndicator, w
ebtedness,
erty measue deprived ind each per
results mayy and modeand Pakista
nt and inted dimensioent of whaf the share
PovertyIndi
47.80%
(0.0006)
verty
26.67%
18.69%
14.97%
17.38%
11.58%
11.72%
EU‐SILC useMalta, Ro
13, several clude indicat
ms, heating
obtained skal training
which consis
and subje
res are then order to rson or hou
y be reporteerate povertan.
erpret if thens. They maat poverty ie of dimen
ices
48.30%
(0.0006)
25.36%
19.88%
15.31%
16.86%
11.16%
11.42%
rs’ databaseomania, Ice
new indicators for:
g);
kills (e.g., vg, open ted of the l
ective asse
e share of wbe identifiesehold is th
ed for morety level may
ey bear somay be set us, consistennsions expe
48.45%
(0.0005)
25.25%
19.63%
16.43%
16.22%
10.77%
11.70%
e of March eland, Nor
ators have a
via adult elearning, evel of edu
essment of
weighted ined as poor. hereby iden
e than one y be reporte
me resembsing a combncy with suerienced by
48.62%
(0.0006)
25.33%
19.45%
17.92%
15.44%
10.39%
11.48%
2014. way and
appeared
ducation, etc.—an cation);
f income
dicators Cut‐offs tified as
poverty ed, as in
lance to bination ubjective y certain
groups consideunion athe meweighteperson wherea
393. Ponormatpovertygroups conside(such aidentifiestructu
394. Roany senrange orange o
5.5.6
395. It includepreferacombin
5.5.6.1
396. Th
5.5.6.2
397. Thmeasur
of peopleered: if somapproach shasure, this ed dimensiwho is de
as if there ar
overty cut‐tive assessmy assessmeof people
ered: if soms cooking wes a personre of the m
obustness tnsitivity to tof poverty cof plausible
Shouldin
is frequend in a natble to repoe them. He
1 Advan
here are thr
Having a
The priodeprivatiweight oof each m
The relatunder th
2 Disad
here are sres within a
Given theasier to
Ch
e. The accume indicatorhould not bwill also afons, povertprived in are 4 dimens
‐offs may bment of wnts, or obse. The accume indicatowith wood),n as poor – easure, this
tests are althe povertycut‐offs. Emweights and
ncomebe
ntly asked wtional MPI,ort monetaere are some
ntagesof
ree main ad
single head
ority of inons is tranf income camay be impl
tionships beis framewo
dvantages
several disn MPI:
e familiarito release
apter5Multi
uracy or inr clearly idee used. If affect the poty cut‐offs at least onesions, then
be set and what povertservation ouracy or inr clearly id, a union apshould not s will also af
ways to bey line, and ampirically, md poverty c
einclude
whether inc, instead oary povertye considera
including
dvantages o
dline pover
ncome povnsparent than be adjuslicitly equal
etween differk because
sofinclud
sadvantages
y with incoa second
idimensional
136
naccuracy oentifies as d human rigoverty cut‐of 33% or
e dimensioncut‐offs of 2
justified usty is, consif the sharenaccuracy oentifies as pproach – ibe used. If ffect the po
performedalso to highmany MPIs ut‐offs (alw
edinthe
come or coof reportiny separatelyations:
gincomea
of including
rty indicato
verty redurough the sted (if ther).
erent depriall indicato
dingincom
s to inclu
ome and comeasure
PovertyIndi
of compondeprived pehts framewoff. For exar 34% may n, or strictl25% or 26%
sing a comstency wite of dimenof compondeprived pn which ana human ri
overty cut‐o
d here too, hlight comphave been
ways conside
MPI?
onsumption g them sey, although
asadime
income:
r that enco
uction relatrelative were are two s
vations andors are draw
measadi
ding incom
onsumptionthat comp
ices
ent indicateople who mwork underliample, if thbe naturaly more tha% provide th
mbination ofh subjectivnsions expeent indicatpeople whony measuredights frameoff.
in order toparisons whproven to ering standa
poverty meparately. Th Armenia
ension
mpasses ev
tive to reeights. Morseparate m
d income cawn from the
mension
me or con
n poverty mplements th
tors must may not beies the struhere are 3 l as they ran one dimhe same intu
f factors, inve or partierienced bytors must o may not d deprivatioework unde
o make tranich are robbe robust aard errors).
measures shTo date, itand Mexic
verything.
eduction inreover, the easures the
an be studiesame data
nsumption
measures, ithose, rathe
also be poor, a cture of equally‐reflect a mension, uition.
ncluding cipatory y certain also be be poor on at all rlies the
nsparent bust to a across a
hould be t seems o (only)
n other relative e weight
ed easily source.
poverty
t can be er than
Chapter5MultidimensionalPovertyIndices
137
discontinuing or changing a familiar statistic. The press and media have proven able to have the ability to understand and communicate two poverty measures, each having their distinctive contribution, effectively.
Income and consumption are volatile and exhibit seasonality. Yet to improve reporting accuracy often short recall periods are used at least for some items. The effect of this is that the status of each household as deprived or non‐deprived at the time of the survey may not reflect their average status across the past year or other period since the last survey. Monetary measures may be accurate “on average” but not at the household level. Yet the MPI requires each indicator to accurately depict deprivation status at the household level.
The sample designs for different survey types may need to be harmonised, and surveys may need to be extended to include all relevant indicators, without jeopardizing data quality.
If the MPI is being used primarily to design and coordinate social policies, the inclusion of income may be less necessary as the MPI will predominantly monitor the outcomes of a set of policies distinct from income poverty.
For the properties of the MPI to be established, income is measured using an absolute poverty line. If a relative poverty line is used for income (as in Alkire and Apablaza 2016), then the poverty focus and deprivation axioms do not hold. Having a mixture of relative and absolute cut‐offs also is conceptually challenging.
Even if income is included, care must be taken in the design of the measure. For example, in the case of Mexico, it appears that economic and non‐economic aspects of poverty are equally weighted. But in fact, the identification procedure is designed to exclude all persons who are not income poor from having the possibility of being identified as poor. Given the evident mismatches between income poverty and other kinds of poverty, and given that these are in part due to non‐sampling measurement error, this is a potentially problematic identification structure. Persons who are multiply deprived in a set of non‐monetary deprivations should have the possibility to be identified as multidimensionally poor unless there is a very good reason for prohibiting this. This is particular the case if, as in Mexico which uses a human rights framework, non‐monetary deprivations are interpreted as violations of social rights.
398. In the end, the decision of whether to report income or consumption and expenditure poverty separately or inside a multidimensional poverty measure is a particularly important decision. There are pros and cons on both sides. The OPHI/UNDP Global MPI (see Section 5.6) does not include consumption poverty because that variable is not included in the surveys employed, so it is not a feasible option for consideration. However, even if it were available in the data, there are benefits to separating international comparisons, given current controversies regarding the PPP exchange rates used in monetary poverty computations. All countries except for Armenia, Kyrgyzstan and Mexico have opted to keep monetary and MPI measures distinct even when in most cases both measures are developed using the same survey instrument.
Recommendation 26: It is recommended to develop separate and complementary measures for monetary poverty and multidimensional poverty.
Chapter5MultidimensionalPovertyIndices
138
Box5.4 TowardsamultidimensionalpovertyindexinGermany
A multidimensional poverty index for an advanced economy like Germany is proposed by Suppa (2017). Drawing on the capability approach as conceptual framework, the Alkire‐Foster method is applied to the German context.
The increasing interest of academics and policy makers in alternative measures for human well‐being also seems to bring about a consensus about relevant dimensions. The proposed multidimensional poverty index for Germany strongly relies on these insights and recommendations, in particular in Stiglitz et al. (2009), Atkinson et al. (2002), Nussbaum (2001) and OECD (2011). Moreover, the specification also relates to the public debate on poverty and deprivation in Germany. Most indicators included in the proposed poverty measure are already considered as “core indicators” by Germany's official reports on poverty and wealth (e.g., Bundesregierung, 2013). These indicators themselves have been selected based on scholarly advice (Arndt and Volkert, 2007).
The proposal uses data from the German socio‐economic panel (SOEP), a rich multi‐purpose household survey (see Wagner et al., 2007). The poverty index is calculated for three points of time (spanning 2001‐2012) and comprises 6 dimensions: education, health, housing, employment, material deprivation, and social participation. Data for this dimension are information about the frequency of engagement in social activities common in contemporary Germany (e.g., attending cultural or religious events, meeting friends, engaging in voluntary work, helping out friends and neighbours, etc.). An individual is considered deprived in social participation if she reports never meeting friends or to never performing any of the other seven activities.
Income is not included for both conceptual and empirical reasons. Conceptual arguments against a lack‐of‐income dimension rely on potential double‐counting, since dimensions for which income is important (e.g., social participation) are already included. Empirically, income poverty is largely captured by material deprivation indicators. The detailed specification is summarized in the table below.
Dimensions are weighted equally and most indicators are also equally weighted within dimensions. For most analyses a poverty cut‐off of k=33 is used. Many results are, however, robust to the choice of k.
From a policy perspective, the chosen dimensions are sensitive to major economic developments. The period of investigation covers for instance extensive labour market reforms and the financial crisis. The decomposability properties of the Alkire‐Foster method allow for disentangling the effect of these developments on the poor. For example, whereas precarious employment and underemployment rise throughout the decade, unemployment and material deprivation both peak around 2007. While unemployment later falls even below its initial level, material deprivation remains high.
Another important question is whether both measures identify the same individuals as poor. Applying a multidimensional poverty cut‐off of k=33 and an income poverty cut‐off of 60% of the median net household equivalent income implies poverty rates of 11‐13%. However, only 5% of the population are identified as poor by both measures, while 8% are only income poor
Chapter5MultidimensionalPovertyIndices
139
and 5% are only multidimensionally poor. This result is robust to different cut‐offs: generally, less than 50% of the income‐poor are also multidimensionally poor. As both measures substantially disagree on who is poor, different policy implications are to be expected.
Recently, SpiegelOnline, a major news portal in Germany, asked what poverty means in Germany. A tool illustrating different approaches to poverty also presents a slightly simplified version of the specification discussed above (http://www.spiegel.de/wirtschaft/armutsrechner‐bin‐ich‐arm‐a‐1093182.html).
The MPI specification of Suppa (2017) is presented in Table 1:
Table 1 MPI specification
Functioning Deprivation Cut‐off Weight
Education Elementary schooling not completed or elementary schooling completed but no vocational qualification
1/12
Less than 10 books in household 1/12
Housing
The house requires major renovation or is ready for demolition 1/18
Any of the following are absent: bath, shower, kitchen, warm water, toilet 1/18
Overcrowding (less than one room per person) 1/18
Health
Partially or severely disabled 1/18
Reporting 2/4 health issuesb 1/18
Body mass index larger than 30 1/18
Material Deprivation
Reporting 2/4 goods missing for financial reasonsc 1/12
Any of the following are absent: life insurance, pension, owning the house or apartment, other house, financial assets, commercial enterprise, tangible assets
1/12
Social Participation
5/7 activities performed neverd; remaining at most less than monthly 1/12
Never meeting friends 1/12
Employment
Unemployed 1/6
Involuntarily, hours worked < 30 1/18
Precariously employed (incl. temporary work ) 1/18
Notes: a Graduation in Germany is usually achieved after 10 years of schooling. bThe four health issues are (i) a strong limitation when climbing stairs, (ii) a strong limitation for tiring activities, (iii) physical pain occurred always or often during the last 4 weeks, and (iv) the health condition limited always or often socially. cThe four goods asked for are (i) a warm meal, (ii) whether friends are invited for dinner, (iii) whether money is put aside for emergencies, and (iv) whether worn out furniture is replaced. dActivities included are (i) going to the movies, pop music concerts, dancing, disco, etc, (ii) going to cultural events (such as concerts, theatre, lectures), (iii) doing sports yourself, (iv) volunteer work, (v) attending religious events, (vi) helping out friends, relatives or neighbours (vii) involvement in a citizens’ group, political party, local government.
Chapter5MultidimensionalPovertyIndices
140
5.6 Casestudy:TheOPHI/UNDPGlobalMPI
399. Another example following the Alkire‐Foster method is the OPHI/UNDP Global MPI. 42 This index may not be relevant in most UNECE countries because it focuses on acute poverty. However, it may be conceptually useful, as it illustrates the kinds of insights that can emerge from rigorously comparable multidimensional poverty measures.
400. The Global MPI is a measure of acute global poverty developed by the Oxford Poverty and Human Development Initiative (OPHI) with UNDP’s Human Development Report Office (Alkire and Santos, 2010, 2014; UNDP, 2010; Alkire and Robles, 2016).43 Acute poverty is understood as a person’s inability to meet simultaneously minimum internationally comparable standards in indicators related to the MDGs44 and to core functionings. The mathematical structure of the index belongs to the family of measures developed by Alkire and Foster (2007, 2011a; Alkire et al.,2015). In particular, the Global MPI uses indicators that were available for more than 100 developing countries in 2009. The primary data sources are the demographic and health surveys (DHS) and multiple indicator cluster surveys (MICS), with some national or regional datasets also included.
Table5.1 Thedimensions,indicators,deprivationcut‐offsandweightsoftheGlobalMPI
Poverty dimension
Indicator Deprived if… Weight
Education
Years of schooling
No household member aged 10 years or older has completed five years of schooling
1/6
Child school attendance
Any school‐aged child is not attending school up to class 8 1/6
Health
Child mortality
Any child has died in the family in the five‐year period preceding the survey
1/6
Nutrition Any adult aged 70 or younger or any child for whom there is nutritional information is malnourished
1/6
Living standards
Electricity The household has no electricity 1/18
Improved sanitation
The household’s sanitation facility is not improved, or it is improved but shared with other households
1/18
Improved drinking water
The household does not have access to improved drinking water (according to MDG guidelines) or safe drinking water is equal or more than a 30‐minute walk from home, roundtrip
1/18
Flooring Dirt, sand, dung or “other” type of floor 1/18
Cooking fuel The household cooks with dung, wood or charcoal 1/18
42 The global MPI currently is computed by the University of Oxford’s Poverty & Human Development Initiative (OPHI) and UNDP’s Human Development Report Office, and both institutions publish national figures. In addition, OPHI publishes extensive disaggregated data, and hosts an interactive databank so users can create their own maps and infographics. 43 The global MPI is one member of the Alkire and Foster class of multidimensional poverty measures that extends the Foster, Greer and Thorbecke class of poverty measures (2011a,b). Alkire et al. (2015) systematically introduce this measurement methodology and situate it in the field of multidimensional methodologies used for poverty comparisons. 44 A revised Global MPI would naturally reflect core poverty indicators in the SDGs.
Povertdimen
401. Thfor whiMPI useeducatibased oscore isa nestedimensmultidimindicatomove forder to
FigureMoving
402. Apovertyaverage
45 MPI es46 For a moriginal M
ty sion
In
Aow
he 2017 Glch data froes informaton and livinon the deps constructeed weight stion is alsomensionallyors. Figure rom individo compute
5.2 gfromind
s already dy (that is, the intensity
stimations priomore detailed MPI indicators
Ch
ndicator
sset wnership
obal MPI aom 2006 ontion from 10ng standardprivation cued based ontructure: eao equally wy poor thos5.2 takes tdual deprivathe Global
dividualde
discussed, the proporti( ) of thei
or to 2004 aredescription os, see Alkire a
apter5Multi
Deprived i
The househotelephone, bown a car or
ssesses munwards are 0 indicatorsds. Each pet‐off presen a weighteach dimensweighted. Fse people wthe examplations to thMPI for eac
eprivation
he MPI refon of the pir poverty (
e available onof the current nd Santos (20
idimensional
141
f…
old does not obike, motorbikr truck
ultidimensioavailable.45
s which areerson is connted in thed average oion is equaFinally, a pwho are depe of a hyphe deprivatch country.
stothede
lects both population t(i.e., the av
line and in deindicator defi010).
PovertyIndi
own more thake or refrigera
onal povert5 As summa organised nsidered to e table. Theof the deprilly weightedpoverty cutprived in 33othetical pion score o
eprivation
the incidenthat is multverage prop
etailed tables uinitions, see A
ices
n one radio, Tator and does
y for peoparized in Tain three dimbe depriveen, each peivations thed, and eacht‐off of 33.33% or moerson (Graof each unit
score
nce or headtidimensionportion of i
upon requestAlkire and Rob
W
TV, not
le in 103 coable 5.1, themensions:46
d in each inerson’s depey experienh indicator w3.33% identore of the wce) and shot (or house
dcount rationally poor), indicators i
. bles (2016). Fo
Weight
1/18
ountries e Global 6 health, ndicator privation ce using within a tifies as weighted ows the ehold) in
o ( ) of and the n which
or the
Chapter5MultidimensionalPovertyIndices
142
poor people are deprived). The Global MPI is calculated by multiplying the incidence of poverty by the average intensity across the poor ( ).47
403. In addition to identifying those who are MPI poor, additional poverty cut‐offs are used, allowing us to identify those who are “vulnerable to poverty” (people deprived in 20% to 33.33% of weighted indicators), and those who are in “severe poverty” (those deprived in 50% or more of the weighted indicators).
404. The Global MPI has been published by UNDP in every Human Development Report since 2010. From 2015 it is being updated twice per year.
Box5.5 AnMPIforKyrgyzstan
An MPI was adapted to Kyrgyzstan (for 2006‐2010) using the Alkire‐Foster method. Poverty was conceptualized as a state in which established social norms and standards influence what constitutes a decent life. People whose levels of goods and services fall below these norms are considered poor. Calculations were based on official household budget and labour force survey data.
Eight indicators belonging to four dimensions (health and nutrition, education and employment, quality of housing, and financial insecurity) were used. All indicators were weighted equally, and while the focused on consumption levels, they were more nuanced than simply income or basic necessities. A household was considered poor if it was simultaneously deprived in at least two indicators.
Table 1 Dimensions and indicators used to identify the poor
Dimensions Indicators Deprived if…
Health and nutrition
Quality of food Consuming less than 2100 kcal daily
Access to healthcare services Unable to get medical care
Education and employment
School enrolment or number of unemployed adults
School‐age children not engaged in education or unemployed adults
Number of people who dropped out of the educational system without completion
Did not reach the required level of education
Quality of housing
Lack of access to clean drinking water Open‐air source of drinking water
Lack of toilet facilities or sewage No toilet facilities
Financial insecurity
Levels of relative poverty Relative poverty
Presence of debt that exceeds 30% of income
Debt exceeding 30% of income
47 The MPI can be equivalently computed as the mean of the censored deprivation matrix, times the number of indicators (here, 10). See Alkire et al. (2015), Chapter 5.
Chapter5MultidimensionalPovertyIndices
143
While the MPI was considerably higher in rural areas than in urban areas during 2006‐2010, it was generally falling in rural areas but unstable in urban areas.
Figure 1 Tendencies of levels of multidimensional poverty in urban and rural areas, 2006‐2010
Source: UNDP Human Development Report of the Kyrgyz Republic (2012).
5.7 Casestudy:Colombia’snationalMPI–Structureandpolicyapplications
405. Colombia’s MPI was launched in 2011 based on the priorities established in the National Development Plan (NDP) of 2011. It uses the Alkire‐Foster method to generate the MPI and the associated set of sub‐ and partial indices. The household is taken as the unit of identification, not only for reasons of data availability but also because of the desire to recognise the importance of the household and incentivize caring and sharing across household members. It defines five equally weighted dimensions, and 15 indicators which are equally weighted within each dimension (Figure 5.3). The weights and poverty threshold are justified both normatively and by reference to subjective poverty assessments, and to the number of deprivations experienced by persons who are, and are not, income poor. The household survey that provides information for the index is fielded annually and the MPI is updated annually, with the data and computational algorithms being publicly available online.
406. Colombia’s MPI has proved to be a powerful tool for informing specific policy actions against poverty and tracking its progress, as it can be broken down to reveal the contribution of each indicator to changes in overall poverty levels and to each of the regions and sociodemographic groups in Colombia. Colombia also developed detailed poverty maps using census data for 11 of the 15 indicators, which provides information to local actors. The MPI directly informs the Families in Action programme that assigns cash transfers to households who improve their educational achievements; it also gives access to benefits
0
0.02
0.04
0.06
0.08
0.1
2006 2007 2008 2009 2010
Rural Urban
MPI
under taccountNationainformsprojectsdisplayeoutside
FigureDimen
So
the Unidos tability in tal Developms targeting s and activied on it, toge governme
5.3 nsions,Var
ource: Angulo
Ch
governmenhe special ment Plan of resourceities by thegether withnt to play th
riablesand
, Pardo and D
apter5Multi
nt programministerial are on tra
es. A social private sec cases of suheir part in
dWeights
Díaz (2011).
idimensional
144
me. Importcabinet rouack. At a gmap at thector and otuccess and ofighting po
fortheMP
PovertyIndi
tantly, it is undtable thgeographicae municipalther non‐goother informverty in all
PIColombi
ices
the basis fhat ensuresally aggreg level is upovernmentamation, in oits dimensio
ia
for monitors the targetgated level,ploaded onlal organisatorder to urgons.
ring and ts in the , it also ine, and ions are ge actors
Chapter5MultidimensionalPovertyIndices
145
Box5.6 LessonsfromColombia’sNationalMPIs
Angulo (2016) describes the general process through which Colombia has designed and implemented its national MPI (C‐MPI). In addition to its statistical dimensions, ensuring that the MPI could be a useful policy tool has meant that this process has also required political coordination.
Because the most significant tools in Colombia’s poverty reduction strategy focus on the household rather than on the individual, the unit of identification was chosen to be the household. Three stakeholders have been involved in the building, dissemination, and application stages of the national MPI: the National Planning Department (DNP), the National Statistics Office (DANE), and the Department for Social Prosperity (DPS). While DANE is the main actor in the dissemination process, the application process at the national level is carried out by the DNP and the DPS. The DNP focuses on monitoring the National Development Plan and public policy design, while the DPS tends to use the C‐MPI either as a targeting tool or for designing and operating social programs.
Table 1 Examples of Applications using C‐MPI
Application Description National roundtable to reduce poverty and inequality
Use of C‐MPI in a high‐level committee for monitoring the national poverty and inequality reduction strategy.
Geographic targeting tool for social programmes
‐ A criterion to introduce geographic differentiation in the conditional cash transfer programme (Program Mas Familias en Accion). ‐ A diagnostic tool for regional development plans elaborated by the DNP and local governments. ‐ A criterion to distribute the overall number of beneficiaries per municipality in several programmes from the DPS.
Social map A geographic tool to encourage public‐private partnership to reduce poverty and inequality and improve the quality of life.
Criteria for graduation from the Colombian safety net to overcome extreme poverty (Unidos)
The C‐MPI and the extreme poverty line are two criteria to graduate households from the safety net Unidos. In this case, the C‐MPI has to be estimated using beneficiary surveys.
Definition of policy combinations to reduce multidimensional poverty and to consolidate the expansion of the middle class
‐ Use of the C‐MPI to identify the most frequent deprivation combinations in order to design public policy and social programmes. ‐ Use of the C‐MPI, in combination with the World Bank’s income methodology, to measure the middle class. The DPS is designing a public policy agenda to foster the consolidation of the middle class in the country.
Source: Angulo (2016).
Chapter5MultidimensionalPovertyIndices
146
5.8 Casestudy:TheMexicanMPI
407. The design of Mexico’s MPI began in 2000 and culminated with its launch in December 2009. It was the first national poverty measures to reflect multiple dimensions along with income. Mexico’s MPI is estimated by the Council for the Evaluation of Social Policy (CONEVAL), which is autonomous body of the Mexican executive. The Mexican index is today one of two such official MPI measures that includes income as a component (the Armenian index is the other one). In the Mexican case, income is weighted at 50%.
408. The national household survey that provides the data for Mexico’s MPI is fielded every two years. The dimensions are defined by the General Law for Social Development (LGDS) based on social rights guaranteed by the National Constitution. The unit of identification is the individual person, so the index can be disaggregated by gender and age. Mexico’s MPI is defined in the economic wellbeing space and in the social rights space. Economic wellbeing is gauged according to national income poverty lines for urban and rural areas. It uses the food poverty line for extreme poverty and the basic needs poverty line for moderate poverty. The social rights space contains six social rights:
a) Deprivation in educational occurs when individuals aged three and above lack the mandatory basic education that prevails for their cohort.
b) Deprivation in access to health services identifies individuals who are not enrolled in or entitled to any mechanism of health protection either public or private.
c) Deprivation in social security for economically active individuals occurs if they do not enjoy the benefits established in the law or are not voluntarily enrolled in social security or retirement investment plan. For those out of the labour force, deprivation in social security occurs when they (or their relatives) cannot benefit from a retirement programme or pension (either voluntary or universal pension system).
d) Deprivation in housing occurs when either the ratio of individuals per room is greater than 2.5; or when the dwelling has a dirt floor or is made of cardboard, metal or asbestos sheets, waste, mud, daub, wattle, reed, bamboo, or palm wood.
e) Deprivation in access to basic services in the dwelling occurs if water is taken from an unprotected or shared source, if drainage is non‐existent or connects to unprotected disposal, if there is no electricity, or if wood or coal are used for cooking inside the dwelling with no chimney.
f) Deprivations in access to food occur in the presence of moderate or severe food insecurity according to FAO (2006).
409. A person is identified as multidimensionally poor if she is deprived in economic wellbeing according to the basic needs poverty line, and is deprived in one or more areas of social rights. Hence, income and social rights have an effective weight of 50%, and each of the social rights has an equal weight of (1/12). A person is in extreme poverty if she is deprived according to a more extreme (food) income poverty line and in at least three social rights. CONEVAL’s report on multidimensional poverty presents the headcount ratio of multidimensional poverty (H), and also the average number of social deprivations among the poor. A modified form of a multidimensional poverty index is reported, which is H times the proportion of social rights in which poor people, on average, are deprived (not including
incomemultidim
FigurePovertDevelo
410. Thidentityreportsprogramtool in Hunger
5.9 A
5.9.1
5.9.1.1
411. A whethenationaindicatoArmeniHondur
because mensionally
5.4 tyidentificopmentPo
Source: Con
he MPI any, and other. Mexico’s mmes and sthe creatio
r and the un
Assessm
Advanta
1 Headl
headline er poverty al or regionor, as a sisa, Bhutan, ras, Pakistan
Ch
if a persoy poor).
cationbytolicy(CONE
neval (2010), u
nd associatr salient chaMPI is us
sectors. In aon of two niversal pen
mentoft
agesofth
lineortie
MPI enable(defined anal MPI coster measuChile, Colon, and Pana
apter5Multi
on is not
theNationEVAL)
used with per
ed statisticaracteristicsed by the addition, themajor policnsion system
heMPI
heMPI
er1pover
es compariacross multould be reure. This is ombia, Costama.
idimensional
147
income d
alCouncil
mission.
cs are decs, and are pCabinet ae Mexican Mcies againstm.
rtyindica
sons acrostiple deprivported aloincreasingta Rica, Do
PovertyIndi
deprived, b
fortheEv
omposed bpresented ins a tool foMPI served t poverty:
ator
s regions avations) hangside thely occurrinominican Re
ices
by definitio
valuationo
by state, gn national aor policy coas a fundathe Nation
and trendss increasede monetaryg nationallepublic, Ecu
on, they a
ofSocial
gender, inds well as staoordinationmental infoal Crusade
that can d or decrey poverty hy—for exauador, El S
are not
digenous ate level n across ormative against
indicate eased. A headline mple in alvador,
Chapter5MultidimensionalPovertyIndices
148
5.9.1.2 Jointdistributionofdeprivations
412. For human development in the context of the SDGs, four criteria are particularly important:
1. multidimensional analysis 2. decomposition (or disaggregation) by regions and population groups 3. interlinkages across dimensions (joint distribution), and 4. weights and robustness
413. The MPI illuminates the overlapping disadvantages poor people experience. It is built from unit data structured to define deprivations for each chosen indicator and person. The deprivation profiles depict the 0‐1 vector of deprivations each person does or does not experience. It uses these (weighted) vectors to identify who is poor, to aggregate information on poverty into a headline measure, and to generate the MPI as well as the incidence, intensity, and indicator composition of poverty. Because of its order of aggregation—first across indicators for each person and then across the population—the MPI captures interconnections between different deprivations for the same person. In this way, the MPI builds upon the counting traditions widely used in Latin America and Europe (Atkinson 2003, 2016). Dashboards and standard composite indices do not capture the joint distribution of deprivations, because they first aggregate information about one deprivation across all units.
5.9.1.3 Informsintegrated,multisectoralpolicies
414. In addition to SDG1, reducing poverty in all its dimensions is a crosscutting goal in the SDGs. Because an MPI incorporates multiple dimensions, it can promote integrated and collaborative policies across a subset of SDG indicators, while prioritizing the poor. According to a July 2015 UNGA document the SDGs are providing “a stronger incentive than in the past for cross‐sector, integrated and collaborative work. Similarly, to evaluate progress under the sustainable development goals, it will be necessary to look at multiple goals concurrently and in an integrated fashion.” In terms of core poverty indicators, an MPI is a tool satisfying the call of the SDGs to “facilitate integration and policy coherence across sectors”.48 At the national level, this has already been a key attraction of the MPI, which animates Mexico’s Crusade against Hunger and Colombia’s Poverty Roundtable, as discussed above.
48 The Economic and Social Council of the UN GA (2015). Available from http://www.un.org/ga/search/view_doc.asp?symbol=A/70/75&Lang=E. This document observes that “Insufficient understanding of and accounting for trade‐offs, interlinkages, synergies and benefits across sectors have at times resulted in incoherent policies, adverse impacts of some sector‐specific development policies and, ultimately, diverging outcomes and trends across broad objectives for sustainable development.” It recognises the need for “United Nations agencies, funds and programmes concerned with a specific goal (e.g., education, health, economic growth)” to take into account targets that refer to other goals”.
Chapter5MultidimensionalPovertyIndices
149
5.9.1.4 Datarequirements
415. Because the MPI has multiple indicators it is often mistakenly presumed that the time and cost of surveys is higher. Actually, the reverse is the case. While the exact proportions depend on the variables included (particularly whether monetary poverty is included, and what kinds of employment questions are addressed), ordinarily MPIs draw on less than 10% of the number of question contained in monetary poverty measures, and take a fraction of the time and cost to implement. MPIs are therefore appropriate for surveys with higher levels of disaggregation. A number of countries have also implemented a limited MPI from census data (Mexico, South Africa, Tunisia, Colombia, and Bhutan).
5.9.1.5 Identifies somewhat different set of the poor than monetarymeasures
416. AN MPI cannot be assumed to identify the same persons as poor nor to proxy the level or the trend of income poverty measures. Many studies have documented the mismatch between non‐monetary deprivations and monetary poverty. This mismatch is also evident between MPIs and monetary poverty measures. For example, in Chile, 14.4% of people are income poor; 20.4% are MPI poor, but only 5.5% are poor in both national measures (Ministerio de Desarrollo Social ‐ Gobierno de Chile, 2015). A study of moderate multidimensional poverty in 17 Latin American countries over time suggests that a significant proportion of the populations are not income poor yet are multidimensionally poor (Santos et al. 2015). Using both income and MPI measures provides a more accurate picture of poverty. Reductions in multidimensional poverty also may not match monetary poverty trends nationally or sub‐nationally. For example, whereas absolute reductions in monetary poverty rates in initially poorer states were faster in India between 1993‐1994 and 2004‐2005, reductions in the MPI rates were slower in initially poorer states during this period (Alkire and Seth 2015).
5.9.1.6 Reflectsamultidimensionalsituationnosingleindicatorproxies
417. Empirical studies also show limited overlaps between deprivations in different indicators. Deprivation in one indicator does not necessarily proxy deprivations in others. For example, the next table shows the deprivation rates of 10 indicators across 101 countries in the second row and second column. Table 5.2 also shows at its centre the proportion of population that showed coupled deprivations in any two given of the 10 indicators. We can point out that although the levels of the two education indicators are very similar (18.4% and 19.9%); their overlap is relatively low, with only 8% experiencing both deprivations. Such a mismatch, which occurs in many indicator pairs, suggests the value of looking at a set of simultaneous deprivations together in order to distinguish those who are deprived in a larger set of indicators from those who are not deprived or deprived in a lesser set.
Chapter5MultidimensionalPovertyIndices
150
Table5.2 Averagedeprivationinpair‐wiseindicatorsacross101developingcountries
Years of
schoo- ling
School atten--dance
Child mortality
Nutri- tion
Electrici-ty
Sani- tation
Drinking water Floor Cooking
fuel
Population deprived in each indicator (in %)
14 14 17 27 22 40 26 27 53
Percentage of population simultaneously deprived in both the column and row indicators Years of schooling 14 School attend-ance 14
5
Child mortality 17
4 5
Nutrition 27 5 6 7 Electricity 22 8 7 8 9 Sanitation 40 10 10 11 15 19 Drinking water 26 5 5 5 8 10 13 Floor 27 8 8 9 12 17 22 9 Cooking fuel 53 12 12 14 19 21 33 19 25 Assets 23 8 7 7 10 14 19 8 16 21
Source: Alkire and Robles (2016).
5.9.1.7 Communicationanddatavisualization
418. The MPI is the product of two easy‐to‐understand and intuitive partial indices. The headcount ratio (H) can be easily explained to journalists, who are already familiar with this idea from monetary measures. And the new partial index of intensity (A)—the percentage of deprivations that poor people in that country face at the same time—creates powerful properties yet also ties the poverty measure back to human lives and experiences. Data visualization examples include maps, poverty composition graphics, bubble charts of incidence and intensity, and so on.
5.9.1.8 Decompositionbypopulationsubgroups
419. Because the MPI is additive and decomposable, and because the data it uses are directly comparable across populations, the MPI, headcount ratio (percentage of people who are poor), intensity (average deprivation score among poor people), as well as all indicator levels and trends can be disaggregated by any subgroup for which the data are representative, such as subnational region, ethnic group, age group, or other social categories. This supports the SDG goal of “leaving no one behind” and seeing whether the poorest groups are catching up over time. For example, the Global MPI mentioned below has been disaggregated for 1468 subnational regions, by rural‐urban areas for all except two countries, by age group (Vaz, 2014; and Vaz, forthcoming 2015), sex, and, for some
Chapter5MultidimensionalPovertyIndices
151
countries, by additional variables such as ethnicity, caste, sex of the household head, and disability status (Alkire and Seth, 2015). In the 2017 Global MPI tables, 984 subnational regions are reported. All low‐income countries, 37 of the 39 Sub‐Saharan African countries, 9 of the 10 East Asian countries, 14 of the 18 Latin American countries, and all South Asian countries covered can be disaggregated sub‐nationally, for example. All national MPIs are disaggregated by relevant groups (geographic, rural‐urban, indigenous ethnicity, etc.); Africa, Mexico, and Colombia for example build national MPIs from census data to obtain poverty maps directly.
Box5.7 Subgroupdecompositionanddimensionalcontribution—Pakistan
In 2016, Pakistan launched its national MPI. The Pakistani Ministry of Planning, Development, and Reform (with assistance from UNDP, OPHI, and the Pakistan Bureau of Statistics) built Pakistan’s national MPI after a series of consultations with key national stakeholders. The measure took the functional form given in the Table 1 below.
Table 1 Pakistan’s National MPI
Dimension Indicator Deprivation Cut‐off Weights Education Years of
schooling Deprived if no man OR woman in the households above 10 years of age has completed 5 years of schooling
1/6 = 16.67%
Child school attendance
Deprived if any school‐aged child is not attending school (between 6 and 11 years of age)
1/8 = 12.5%
School quality Deprived if any child is not going to school because of quality issues (not enough teachers, schools are far away, too costly, no male/female teacher, substandard schools), or is attending school but remains dissatisfied with the service
1/24 = 4.17%
Health Access to health facilities/ clinics/Basic Health Units (BHU)
Deprived if health facilities are not used at all. Or are only used once in a while, because of access constraints (too far away, too costly, unsuitable, lack of tools/staff, not enough facilities)
1/6 = 16.67%
Immunisation Deprived if any child under the age of 5 is not fully immunised according to the vaccinations calendar (households with no children under 5 are considered non‐deprived).
1/18 = 5.56%
Ante‐natal care Deprived if any woman in the household who has given birth in the last 3 years did not receive ante‐natal check‐ups (households with no woman who has given birth are considered non‐deprived).
1/18 = 5.56%
Assisted delivery
Deprived if any woman in the household has given birth in the last 3 years attended by untrained personnel (family member, friend, traditional birth attendant, etc.) or in an inappropriate facility (home, other) (households with no woman who has given birth are considered non‐deprived).
1/18 = 5.56%
Standard of Living
Water Deprived if household has no access to an improved source of water according to MDG standards, considering distance (less than a 30 min return trip): tap water, hand pump, motor pump, protected well, mineral water
1/21 = 4.76%
Chapter5MultidimensionalPovertyIndices
152
Dimension Indicator Deprivation Cut‐off Weights Sanitation Deprived if household has no access to adequate sanitation
according to MDG standards: flush system (sewerage, septic tank and drain), privy seat
1/21 = 4.76%
Walls Deprived if the household has unimproved walls (mud, uncooked/mud bricks, wood/bamboo, other)
1/42 = 2.38%
Overcrowding Deprived if the household is overcrowded (4 or more people per room)
1/42 = 2.38%
Electricity Deprived if the household has no access to electricity 1/21 = 4.76% Cooking fuel Deprived if the household uses solid cooking fuels for cooking
(wood, dung cakes, crop residue, coal/charcoal, other) 1/21 = 4.76%
Assets Deprived if the household does not have more than two small assets (radio, TV, iron, fan, sewing machine, video cassette player, chair, watch, air cooler, bicycle) OR no large asset (refrigerator, air conditioner, tractor, computer, motorcycle), AND has no car.
1/21 = 4.76%
Land and livestock (only for rural areas)
Deprived if household is deprived in land AND deprived in livestock, i.e. a) Deprived in land: the household has less than 2.25 acres
of non‐irrigated land AND less than 1.125 acres of irrigated land
b) Deprived in livestock: the household has less than 2 cattle, fewer than 3 sheep/goats, fewer than 5 chickens AND no animal for transportation (urban households are considered non‐deprived)
1/21 = 4.76%
Source: Multidimensional Poverty in Pakistan (Official Report, 2016).
One of the special features of this MPI is its high‐resolution lens. It provides an aggregate headline figure, but also allows for subgroup decomposition and dimensional breakdown. By dividing the population into mutually exclusive and collectively exhaustive sub‐groups, the overall MPI can be expressed as an average of the sub‐group MPIs. It is also useful to observe the contribution of each indicator to overall poverty. Pakistan’s MPI was decomposed by rural and urban regions, by provinces, and by districts. Percentage contribution of each indicator was also computed.
It was found that a lot of variation exists at the regional as well as the district level. The map below shows the variation in the headcount ratio of those who are multi‐dimensionally poor across the various districts. This shows that while the headline figure is important and informative, sub‐group decomposition allows policy to be targeted towards poor regions directly. This shows where poor people are. An investigation into the contribution of each dimension to overall poverty shows how people are poor. Combined with information from the headcount ratio (how many people are poor) and the intensity of poverty (how poor they are), the high‐resolution lens of MPI becomes a very valuable tool for policymaking.
To illustrate the dimensional contribution, Pakistan has investigated the percentage contribution of each indicator to overall poverty, both at the national level and at the rural/urban levels. The results are presented in the Figure 1 below.
Figure 1Percent
S
5.9.1.9
420. Ain each and ovtraditioand pounpack for morentry ointensitcut‐offscontrib
1 tage contrib
Source: Multi
9 Coord
ny MPI can indicator ter time. Suonal dashboor (and is athe MPI is re than oneof the Globty A (povers of 20% ution (weig
Ch
bution of e
dimensional P
dinatedda
be publishhat composuch informoard of nonalso reportefocused one poverty libal MPI, foty cut‐off 3and 50%, ghted) of e
apter5Multi
ach indicat
Poverty in Pak
ashboard
ed with a cses it, and ation is esn‐monetaryed), the co the poor. Tine or cut‐oor example,33%), dimencensored
each indicat
idimensional
153
or to MPI, b
kistan (Officia
dandtable
complementhow the cossential to y indicators ordinated dThe MPI andoff. For eac, online tansional conheadcounttor to over
PovertyIndi
by national
l Report, 2016
es
tary dashboomposition guide integreflects de
dashboard d its associach national,bles reportntributions, t ratio for rall poverty
ices
l and rural/
6).
oard showinof poverty grated polieprivations of consisteated statist rural‐urbat the headheadcounteach ind
y, and unce
/urban
ng the deprvaries by suicies. Wherof all peopnt sub‐indiics can be rn, and subcount ratiot ratios for icator, perensored hea
rivations ubgroup reas the ple, rich ces that reported national o H and poverty rcentage adcount
ratios favailabldestitutamong similar formula
5.9.1.1
421. Adirect mDirect mneeds. access tguarantmeasurcountriand assdefinitioevolving
5.9.1.1
422. Thand trepermitsacross MPI levover tim
5.9.2
5.9.2.1
423. JupovertymeasurinequalMPI). Emeasurindicato
49 Alkire aChapter 8level and
for each indle online (ote (using a the poor atables areation and m
10 Cross
longside indmeasures omethods veIndirect meto the resotee them. re is that ses and do sociated haons (safe wg SDG indic
11 Statis
he MPI, thends are pubs assessmencountries avels, and deme, either n
Disadva
1 Confu
ust as county, so too thre of wellblity (such aach measure human dors of well‐b
and Santos (28 of Alkire et d trend.
Ch
dicator. Staothers by reset of destas well as issued, an
monitoring.
s‐national
direct moneof poverty, rify whetheethods suchurces whichAn additiostandardizednot requirearmonizatio
water, adequcators.
sticalinfe
e headcounblished withnts of statiand subnatieprivation lenationally or
antages
usionwith
tries normae suite of hbeing (suchas the ineqre is differeevelopmenbeing or ine
014) compareal. (2015) cov
apter5Multi
ndard erroequest). Thetitution depany drop ind the indi
lcompari
etary measuwhich referer people ach as an inch might enaonal advantd assessmee adjustmeon of sociauate sanitat
erence
nt ration (Hh their respestically signional groupevels in eacr by disaggr
hcomposi
ally measurehuman deveh as the huality‐adjusent and valnt. But it is iequality.
e results for bvers analytical
idimensional
154
rs and confe tables alsprivation cun the sampicator dash
isons
ures of povrence deprctually achieome or conable them ttage of dirents can bents for PPPal indicatortion, etc.); a
H), the inteective standnificant diffps.49 It also ch componregated gro
iteindica
e income pelopment mhuman devsted HDI), aluable in itsimportant n
ootstrapped astandard erro
PovertyIndi
fidence inteo report thut‐offs for kple retainedhboards are
verty, Amartivations in eve certain nsumption to achieve srect measue comparedP exchange rs have staa regional M
nsity (A), adard errors ferences acenables coent indicatoups.
ators
er capita, inmeasures havelopment and a meas own rightnot to conf
and analyticalors and statist
ices
ervals for nhe subset ofkey indicatod for estime used ext
tya Sen (19functioningfunctioningbudget setsuch functioures for a d directly arates or inandardized MPI may bu
nd consisteand confid
cross poverountries to or, have sig
ncome ineqave a measindex [HDsure of pot, and they use indicato
l standard errtical inference
national figuf the poor wors), and inmations. Naensively fo
81) proposgs or their gs or satisfyt establish ponings—butregional oacross regionflation. Themany depild on these
ent indicatoence intervty levels, inascertain wgnificantly d
quality, andure of eachDI]), a meaverty (suchcan all be ors of pove
rors for the Gle for comparis
ures are who are equality tionally, or policy
es using proxies. y certain people’s t do not r global ons and e MDGs privation e and on
or levels vals. This ncluding whether declined
income h type: a asure of h as the used to
erty with
obal MPI. sons of
Chapter5MultidimensionalPovertyIndices
155
424. The largest single disadvantage of MPI is that, because the methodologies are relatively new, statisticians are not familiar with them. In particular, multidimensional poverty indicators are often incorrectly confused with “composite” indicators which first aggregate across unit data, and then build an overall measure. Examples of these include the OECD Your Better Life Index, the social progress index, and the human development index. Composite indicators have very different properties. They do not reflect the joint distribution of deprivations. They do not identify who is poor. Also, the weights for composite indicators are required to play a much more demanding role as mentioned above, because they generate marginal rates of substitutability between indicators at different levels of achievement. In contrast, the MPI, like a monetary poverty index or like the material deprivation index, is based on unit record data, and aggregates this for each person or household, identifies who is poor, and only subsequently builds a national measure.
5.9.2.2 Rareevents
425. No measure will sufficiently reflect poverty for all social groups in all dimensions. Thus, the MPI will ordinarily be interpreted alongside a small set of indicators. Some indicators, such as those pertaining to rare events, such as maternal mortality or to small populations (the percentage of national political leaders who are women), are not ordinarily included in an MPI because if they are combined with far more frequent deprivations such as lack of sanitation, then they may always appear to have a very low frequency and contribution, and it may be difficult to obtain statistically significant changes.
426. The MPI’s ability to provide an overview of disadvantages of different population groups may be improved in the design phase. For example, many countries include variables pertaining to childhood and youth in their MPI, or else have a separate dimension focused on childhood and youth conditions. In the absence of such an effort, there may be a value in developing a supplementary MPI for children, which is able to highlight the differing challenges faced by children 0‐17 in different cohorts across the society. Bhutan, for example, has chosen to produce an MPI for children aged 0‐17 which has 50% the same indicators as their national MPI, and 50% child‐specific indicators, which are separately defined for each age cohort and, in some cases, gender.
As was mentioned in chapter 3, given the populations currently overlooked by household surveys, including the homeless and institutionalized populations, special studies may need to be performed in order to assess the poverty of certain groups which may have unusually high but invisible levels of poverty (recommendation 3).
Chapter6ChallengesfortheFuture
156
6 ChallengesfortheFuture
427. This Guide focuses on areas where the statistical community has expressed a particular need for further guidance in measuring poverty. It includes recommendations wherever these are warranted by current knowledge and experience. Yet, current knowledge on several key issues is not sufficient for the formulation of clear‐cut recommendations; further work is required. The most important of these issues are briefly summarized below. It is expected that international organizations will coordinate further methodological work on these topics.
6.1 Hardtoreachpopulations
428. In an era of falling response rates, it may seem that all household populations are hard to survey, but some populations present special challenges that make them harder (Tourangeau et al., 2014). When measuring poverty through poverty surveys, it should be recognised that certain categories of people who may be among the most likely to be poor are frequently omitted from the sampling frame since they do not live in private households. The welfare of marginalized groups such as homeless people (including street children), drug users, sex workers, people who are in institutions (including elderly care homes, children’s homes, and mental health institutions); people in temporary accommodation or hostels; prisoners; and refugees in camps or illegal immigrants is notoriously difficult to assess systematically.
429. In most demographic studies where representative household surveys are the gold standard for data collection, such marginalized population segments are lost by definition or, at best, are grossly underrepresented. An additional set of problems concerns the willingness of individuals from marginalized, hard‐to‐reach groups to provide information to survey interviewers, especially about the type of sensitive matters that define their marginality (undocumented migrants), or difficulty in being interviewed (low literacy groups, or where there is a language barrier). It can be argued that the same social characteristics and constraints that hinder access to these individuals may also impair their willingness or ability to answer survey questions. These types of bias led to significant underestimation of poverty rates, for example in EU‐SILC (Nicaise and Schockaert, 2014). It is therefore also important when reporting estimates to inform the user about the potential sources of bias. Box 6.1 provides an additional example of underestimating income inequality due to the varying non‐response of population living in the capital and rural regions in Ukraine.
430. A significant number of children would be excluded from usual survey samples. For example, in many countries in Central and Eastern Europe between 1 and 2 per cent of children are growing up in institutionalised care, and children from poor or excluded backgrounds have been shown to be more likely to be institutionalised (United Nations Children’s Fund Regional Office for CEE/CIS, 2014). Others are involved in seasonal migration or in marginal or excluded communities (e.g., Roma), whose households may not be included in surveys. Children may be omitted from household responses due to the reluctance of the respondent to recognize as household members foster children, trafficked children, or children who are engaged in hazardous child labour.
Chapter6ChallengesfortheFuture
157
431. Some populations are less represented in household surveys because they live in remote areas, enclaves or in families or communities whose presence is illegal and unregistered, for example Roma populations (McDonald and Negrin, 2010). UNICEF has conducted Multiple Indicator Cluster Surveys on Roma populations in Serbia and Montenegro to obtain data on these populations. (See Box 3.2 for the approach developed by UNDP in responding to these challenges.)
432. The challenge remains that surveys are better suited to cover only the easily accessible populations. Covering all the “hard to sample”, “hard to identify”, “hard to find or contact”, “hard to persuade” or “hard to interview” categories of respondents may make a survey complicated and costly. Such a challenge must be met with innovative strategies for sampling, identifying, locating, contacting and interviewing.
Box6.1 HouseholdsurveysinUkraine
A recent study conducted by order of the State Statistics Service of Ukraine (SSSU) by experts of the Academy of Sciences of Ukraine and supported by the World Bank found a systematic downward bias in the income distribution data of household surveys in Ukraine. Non‐response rates for household living conditions surveys during 2010‐2012 in Kiev ‐ Ukraine's capital region ‐ were double the national average, and many times greater than the non‐response rates reported for Ukraine's poorer rural regions (Figure 1). Correcting this bias on base of one of the approaches which are being tested in the SSSU would raise Ukraine's official Gini coefficient for income equality from 23 to 27 (Sarioglo, 2016).
Figure 1 Household living standards survey non‐response rates in Ukraine, 2010‐2012
Source: State Statistics Service of Ukraine.
4
19
48
3
17
46
3
17
50
0
10
20
30
40
50
60
Rural National Kyiv
2010
2011
2012
percentage
Chapter6ChallengesfortheFuture
158
6.2 Imputedrentandhousingcost
433. Evidence suggests that housing represents between 14% and 25% of net adjusted disposable income in OECD countries (OECD, 2014) and up to 30% of the consumption aggregate in developing countries (Balcazar et al., 2014; Deaton and Zaidi, 2002). Omitting housing in the welfare aggregate may mean ignoring substantial differences in quality of life between otherwise similar households, hence resulting in distorted household welfare rankings.
434. Measuring housing value is challenging, both in terms of housing wealth and imputed rents. As with any durable good, the amount that should be captured is the value of the flow of services associated with occupying the dwelling (Balcazar et al., 2014). A good indicator, such as the rental value, may not be available for most households. For instance, in many countries of Eastern Europe and Central Asia, home ownership is prevalent and rental markets are thin. In addition, these markets tend to be composed of dwellings and households with specific characteristics, making it difficult to use actual rental information to impute rental values to the rest of the country. Self‐reported implicit rent may be unreliable because of limited information about rental markets among homeowners. Real estate markets may also move very slowly and long‐time homeowners may not be able to report accurately the sale value of their dwellings, limiting the potential use of such information for imputing rent if collected in surveys. Finally, different countries may use different methodologies to impute rent among homeowners, which complicates comparisons across countries on important social indicators such as poverty or inequality.
435. In light of these issues, improving estimates of the value of housing services could significantly improve welfare measurement in the region and increase international comparability of social indicators. Some efforts have already been undertaken in this direction. For instance, Balcazar et al. (2014) documents the most commonly used methods for rent imputation and provide a basic discussion of its inclusion on poverty and inequality. Cancho and Azevedo (2016) discuss potential sources of bias when using self‐reported implicit rent information. However, more systematic work is necessary to identify:
best practices for rent imputation
guidelines to evaluate when it is better to impute or to ignore the rental information
the actual effect on poverty, inequality and the composition of the poor that the inclusion of imputed rent implies
6.3 Socialtransfersinkind
436. Taking social transfers in kind (STIK) into account in household income and consumption measures is important for comparing rates and experiences of poverty in different countries, as well as for making international comparisons of the level of economic well‐being more broadly. However, because of measurement challenges they are often excluded from the welfare measures used for poverty statistics.
Chapter6ChallengesfortheFuture
159
437. OECD research (Balestra and Sustova, 2017) shows that the majority of countries that contribute to their OECD Income Distribution Database do not produce any estimates of these transfers, with even smaller numbers including them within their income statistics.
438. The types of benefit included within social transfers in kind also vary from country to country, affecting comparability. For example, Tonkin et al. (2014) showed that while both the United Kingdom and Finland include the value of health and education in‐kind transfers in their statistics on income (re)distribution, social housing and public transport subsidies were included in the United Kingdom but not in Finland, and elderly care was included in Finland only.
439. A further challenge to comparability comes from the different methods used. Tonkin et al. (2014) highlight that the value of healthcare services for individuals and households is estimated using an “insurance value” approach in the United Kingdom, but an “actual consumption” approach in Finland. Balestra and Sustova (2017) suggest such methodological differences are widespread, with 30 per cent of countries producing estimates of social transfers in kind using the insurance value method, the same proportion using actual consumption, and 40 per cent using a combination of the two.
440. To address limited data availability, it is recommended that statistical compilers not currently producing estimates of the distribution of social transfers in kind consider including them in income and consumption expenditure statistics. This work should be supported at the international level through guidelines that support countries in adopting a common methodological approach.
6.4 Wealth
441. Income and consumption provide only a partial view of the economic resources that are available to individuals and households. Knowing about the levels of wealth (or debt) that people have is crucial to better understanding their economic well‐being and therefore their experience of poverty.
442. One way of considering how household wealth holdings affect poverty is through asset‐based poverty measures, where asset poverty is defined as an individual having insufficient wealth to meet their basic needs over time. This might be operationalised as having net liquid financial wealth insufficient to cover three months of 60% of median income, taking account of household composition using an appropriate equivalisation scale (e.g. Azpirate, 2008; Tonkin, Serafino and Davies, 2016). Combining such measures with income‐ or consumption‐based poverty measures would allow for distinguishing among the income poor those who have sufficient financial wealth to maintain their material living standards at an appropriate level at least for a short time (income‐poor only), versus those who lack such a buffer (asset and income poor).
443. It is also possible to identify those whose income is currently above the poverty line, but who lack the assets to protect themselves from a sudden fall in their income through, for example, losing their job (asset‐poor only). Being able to identify these different groups within a population would help policymakers to target interventions to reduce poverty.
Chapter6ChallengesfortheFuture
160
444. One of the primary reasons why such measures are not more widely used at present is the limited availability of wealth distribution data, particularly for the same people who are included in the income or consumption poverty statistics. Collecting information on income and/or expenditure and wealth simultaneously is challenging due to the length of questionnaires required. Such data is primarily collected via survey, resulting in high respondent burdens and therefore potentially low response rates.
445. Integrated data collection for income, expenditure, and wealth has been successfully undertaken by some statistical agencies—for example by the Australian Bureau of Statistics through their integrated Survey of Income and Housing (SIH) and Household Expenditure Survey (HES). Increasing use of administrative data by producers would further increase opportunities for such integrated models.
446. However, for many producers, the best opportunities for integrated analysis of income, consumption, and wealth lie in the use of statistical or synthetic matching techniques, combining multiple data sources for different sets of individuals based on common, harmonised matching variables (see Tonkin et al. (2016) for an example). It is therefore recommended that work should be undertaken to develop the use of these techniques in this area of statistics, including the development of guidance and identification of best practices where appropriate.
6.5 Comparablewelfareaggregates
447. To examine poverty and inequality, one needs a measure of material well‐being. Ideally, this measure should correspond as closely as possible to the way a person experiences his or her standard of living. It is natural to think that a person’s standard of living, or material wellbeing, is a function of all goods and services consumed by that person. Economic theory allows one to rank levels of well‐being using the cost (monetary value) of the consumption bundle consumed in a given period. In theory, any welfare measure should include all of the factors (including health, leisure, social capital, and other desiderata) that contribute to welfare. In practice, however, because of measurement and valuation difficulties, the focus in microdata analysis is only on material well‐being, using information on consumption of goods and services by a household. Even such “simple” measures are, in practice, quite complicated to capture well, and there is debate as to whether income or consumption is the preferable measure (see Deaton and Zaidi (2002)).
448. Regardless of the measure chosen, it should be comprehensive and no aspect of income or consumption should be omitted. Within and across countries, the components of income or consumption are heterogeneous due to differences between surveys and the availability of items across countries. Thus, more systematic work needs to be done in terms of
constructing comparable welfares as well as its components
developing sets of key questions that are common for the countries in the region
the impact of different sets of components in the welfare measure on the estimates of poverty and inequality
Chapter6ChallengesfortheFuture
161
6.6 Comparabilityofmultidimensionalpovertymeasures
449. This Guide recommends countries to report a Multidimensional Poverty Index against target 1.2 of the SDGs (recommendation 24). In the short term, countries can report existing national MPIs (Armenia, Kazakhstan, Kyrgyzstan, and Tajikistan), or the value of their global MPI published by UNDP. The guide also recommends that it is desirable to include the dimensions of living standards, services, health, education, work and the lived environment in the national MPI (recommendation 23).
450. In the medium term perspective, countries participating in the Conference of European Statisticians could engage in a process to enhance comparability of their measures of multidimensional poverty and their MPIs. For example, for groups of countries or sub‐regions, a fully harmonized MPI could be developed and agreed upon. Among all countries participating in the Conference of European Statisticians, such agreement could pertain to core indicators for measuring the six core dimensions recommended in this Guide (living standards, services, health, education, work and the lived environment). International exchange and work on this topic would benefit countries in producing the SDG indicator 1.2.2 “Proportion of men, women and children of all ages living in poverty in all its dimensions according to national definitions”.
6.7 Individuallevelpovertymeasures
451. Traditional poverty measures usually take the household as the unit of observation. The focus on aggregate units rather than on individuals is based on the assumption that resources are shared equally within the household. Using a collective unit like the household has two consequences: no one can be counted as poor in households above the poverty line; and all household members are assumed to be equally poor.
452. Both quantitative and qualitative studies have found that household‐level variables are not necessarily optimal predictors of individual well‐being and poverty status, as they ignore gender and other intra‐household inequalities (such as those based on age), as well as the possibility that non‐earners may be poorer than other earning adults in the same household. For instance, research has suggested that, particularly in low‐income households, the assumption of income sharing does not always hold as men sometimes benefit at the expense of women from shared household income (Department for Work and Pensions, 2004). However, assuming no pooling of income is equally problematic.
453. Opening the household black box and producing poverty measures at the individual level requires the use of variables that specifically aim to depict the processes involved in the acquisition and expenditure of resources within the family: the entry of resources into the household; how resources are allocated and controlled; and how resources are expended (Daly, 1992). Unfortunately, such variables are quite rare in household surveys.
454. The development of poverty measures at individual level is still in its early stages, and further research is needed before making recommendations. In the meanwhile, the researcher must make careful methodological choices about the unit of measurement and analysis, and be aware of the implications of these choices in her research. It is also advised
Chapter6ChallengesfortheFuture
162
that, in the identification of the poor, different poverty measures, based on information collected both at household and individual levels, are combined. Finally, adequate and comparable datasets are needed that allow for the study of intra‐household inequalities in poverty by providing information on the acquisition and expenditure of economic resources within the household.
6.8 Spatial differences within countries with regard toconsumptionandincomepoverty
455. Surveys were often collected over a long period of time and across many administrative areas within the countries. In this case, it is necessary to adjust for changes in prices over time using quarterly or monthly consumer price index series. Regional price differences can also cause the same bundle of goods to be more expensive in one region than in another. However, differences in expenditure caused by these regional price differences are not reflected in measured well‐being or welfare. Thus, these regional price differences need to be corrected. For example, the spatial price adjustment for the supplemental poverty measure in the United States is based on median rents of a specific type of rental unit, because food prices do not vary that much for the United States while rents do (Renwick et al, 2014; Renwick and Fox, 2016).
456. The Paasche price index (Deaton and Zaidi 2002) offers a reliable way to measure the spatial price differences. Because non‐food prices are usually not available for countries and unit values for non‐food items are normally not collected by household surveys, the spatial price deflator is based entirely on difference on food prices. More work needs to be done to identify:
the most reliable method for spatial price deflators for income and consumption welfare, and
best practices and methods to include non‐food prices given the heterogeneity of items within and across countries.
6.9 Subjectivepoverty
457. Measures based on subjective perceptions of poverty could contribute greatly to the understanding of poverty and complement the other measures examined in this Guide. Subjective measures can inter alia help to identify weights for different dimensions of welfare and determine the social threshold below which people tend to think they are poor (Ravallion, 2012). Moreover, diagnosis of perception of poverty (self‐assessment of situation, opinions on a scope of phenomenon in a local community or in the country) shows to what extent assessment based on the so‐called objective measures is in line with the opinions of society regarding this problem.
458. While no measures of subjective poverty have been explicitly agreed to internationally, the EU‐SILC indicator of “making ends meet” has been used in comparative studies (e.g., Guagnano et al., 2013; Noll and Weick, 2014). Further examples include the
Chapter6ChallengesfortheFuture
163
Poverty and Social Exclusion Survey in the United Kingdom, 50 the Social Cohesion Survey in Poland (Bieńkuńska and Piasecki, 2016), and in different surveys in the United States (Garner, 2003). The module on the social perception of poverty was also included to the Eurobarometer surveys conducted in 2009 and 2010 (Special Eurobarometer 355, European Commission). Moreover, in May 2015, the OECD launched “Compare your income”51, a pioneering online interactive tool that aims to raise awareness among people in OECD countries on how economic resources are distributed, and which asks users to provide their own assessment of the minimal income that a household as they own would need to avoid poverty.
459. The 2015 UNECE survey on methods of poverty measurement in official statistics showed great variation across countries in the understanding and measurement of subjective poverty. The questions used in surveys used can be grouped as follows:
Ability to meet various needs—financial restrictions faced by the household
Considering oneself as poor—individual self‐assessment
Income necessary to make ends meet—households’ minimum perceived needs
460. The obtained estimates vary significantly due to different methods and cultural perceptions of well‐being, including poverty. Large variations have also been observed within countries by age, gender, and region.
461. The challenges and progress made in obtaining comparable estimates for subjective measures of well‐being (OECD, 2013c) and health (UNECE, 2012c; Robine et al., 2013) are well documented. Such progress is also needed for robust internationally comparable measures of subjective poverty. This can be achieved through the following activities:
Collecting information on and evaluating the results of different international and national experiences
Identifying a small set of measures that could add the most value to the understanding of poverty and could lend themselves to international harmonization
Testing survey questions in different countries and with different population groups
Consolidating the evidence into concrete recommendations for routine data collection in official statistics.
50 See www.poverty.ac.uk. 51 See www.oecd.org/statistics/compare‐your‐income.htm.
164
References
Aaberge, Rolf, Audun Langørgen, and Peter Lindgren (2013). The distributional impact of public services in European countries. Eurostat methodologies and working paper, 2013 edition. Luxembourg: Eurostat. Available from http://ec.europa.eu/eurostat/ documents/3888793/5857249/KS‐RA‐13‐009‐EN.PDF/cd22d557‐f7b7‐47fb‐8841‐9a0fd9084f64.
Alkire, Sabina, James E. Foster, and Maria Emma Santos (2011). Where did identification go? Journal of Economic Inequality, Vol. 9, No. 3, pp. 501–505.
Alkire, Sabina, James. E. Foster, Suman Seth, Maria Emma Santos, Jose. M. Roche, and Paola Ballon (2015). Multidimensional poverty measurement and analysis. Oxford: Oxford University Press.
Alkire, Sabina, and Mauricio Apablaza (2016). Multidimensional Poverty in Europe 2006–2012: Illustrating a Methodology. OPHI Working Paper, No. 74. OPHI, Oxford University. Available from http://www.ophi.org.uk/multidimensional‐poverty‐in‐europe‐2006‐2012‐illustrating‐a‐methodology/.
Alkire, Sabina, and James E. Foster (2007). Counting and multidimensional poverty measures. OPHI Working Paper, No. 7. OPHI, Oxford University.
Alkire, Sabina, and James E. Foster (2011a). Counting and multidimensional poverty measurement. Journal of Public Economics, Vol. 95, No. 7, pp. 476–487.
Alkire, Sabina, and James Foster (2011b). Understandings and misunderstandings of multidimensional poverty measurement. Journal of Economic Inequality, Vol. 9, pp. 289–314.
Alkire, Sabina, and Gisela Robles (2017). Multidimensional Poverty Index Summer 2017: Brief Methodological Note and Results. OPHI Briefing 44. OPHI, Oxford University.
Alkire, Sabina, and Gisela Robles (2016). Measuring multidimensional poverty: Dashboards, Union identification, and the Multidimensional Poverty Index (MPI). OPHI Research in Progress 46a. OPHI, Oxford University. Available from http://www.ophi.org.uk/wp‐content/uploads/OPHIRP046a.pdf
Alkire, Sabina, and Maria Emma Santos (2010). Acute multidimensional poverty: A new index for developing countries. OPHI Working Paper, No. 38. OPHI, Oxford University. Available from https://www.ophi.org.uk/wp‐content/uploads/ophi‐wp38.pdf.
Alkire, Sabina, and Maria Emma Santos (2014). Measuring acute poverty in the developing world: Robustness and scope of the multidimensional poverty index. World Development, Vol. 59, pp. 251–274.
Alkire, Sabina, and Suman Seth (2015). Multidimensional poverty reduction in India between 1999 and 2006: Where and how? World Development, Vol. 72, pp. 93–108.
Allardt, Erik and Hannu Uusitalo (1972). Dimensions of Welfare in a Comparative Study of the Scandinavian Societies. Scandinavian Political Studies, Vol. 7, pp. 9–27.
Amore, Kate, Michael Baker, and Philippa Howden‐Chapman (2011). The ETHOS Definition and Classification of Homelessness. An Analysis, European Journal of Homelessness, vol. 5, No. 2, pp. 19‐37.
References
165
Angulo, Roberto (2016). From Multidimensional Poverty Measurement to Multisector Public Policy for Poverty Reduction: Lessons from the Colombian Case OPHI Working Paper 102. OPHI, Oxford University. Available from http://www.ophi.org.uk/from‐multidimensional‐poverty‐measurement‐to‐multisector‐public‐policy‐for‐poverty‐reduction‐lessons‐from‐the‐colombian‐case.
Asian Development Bank (2008). Research Study on Poverty‐Specific Purchasing Power Parities for Selected Countries in Asia and the Pacific. Manila: Asian Development Bank. Available from http://siteresources.worldbank.org/ICPINT/Resources/270056‐1255977254560/6483625‐1398442720255/Poverty‐Specific‐PPP(ADB2008).pdf.
Atkinson Anthony, Bea Cantillon, Eric Marlier, and Brian Nolan (2002). Social Indicators: The EU and Social Inclusion. Oxford: Oxford University Press.
Atkinson, Anthony Barnes, and Eric Marlier (2010). Income and Living Conditions in Europe. Eurostat, Statistical Books. Luxembourg: Publications Office of the European Union.
Atkinson, Anthony (2003). Multidimensional deprivation: Contrasting social welfare and counting approaches. Journal of Economic Inequality, Vol. 1, No. 1, pp. 51‐65.
Azpitarte, Francisco (2008). Measurement and Identification of Asset‐ Poor Households: A Cross‐National Comparison of Spain and the United Kingdom. Available from http://www.ecineq.org/milano/WP/ECINEQ2008‐105.pdf
Balcazar, Carlos Felipe, Lidia Ceriani, Sergio Olivieri, and Marco Ranzani (2014). Rent Imputation for Welfare Measurement: A Review of Methodologies and Empirical Findings. World Bank Policy Research Working Paper 7103. Available from http://documents.worldbank.org/curated/en/313651468152726645/pdf/WPS7103.pdf.
Balestra, Carlotta, and Sarka Sustova (2017). On the measurement of imputed rent and social transfers in kind in OECD countries. OECD Statistics Working Papers, forthcoming, Paris: OECD Publishing.
Bartels, C., and Stockhausen, M. (2017). “Children's Opportunities in Germany – An Application Using Multidimensional Measures.” German Economic Review, 18(3):327‐376.
Bee, Adam, Bruce D. Meyer, and James X. Sullivan (2015). The Validity of Consumption Data: Are the Consumer Expenditure Interview and Diary Surveys Informative? In Improving the measurement of consumer expenditures, Christopher D. Carrol, Thomas Crossley and John Sablehaus, eds. Chicago: University of Chicago Press.
Beegle, Kathleen G. (2004). Living standards and poverty in Moldova. Washington, D.C.: World Bank Group. Available from http://documents.worldbank.org/curated/en/2015/11/ 25486905/living‐standards‐poverty‐moldova.
Bieńkuńska, Anna, and Tomasz Piasecki (2016). Social perception of poverty and income inequalities in Poland based on the Social Cohesion Survey 2015. UNECE seminar on poverty measurement, Geneva, 12‐13 July 2016, Working paper 14. Available from http://www.unece.org/fileadmin/DAM/stats/documents/ece/ces/ge.15/2016/Sem/WP14_Poland_ENG.pdf.
Bradshaw, Jonathan, and Emese Mayhew (2011). The Measurement of Extreme Poverty in the European Union. European Commision, Directorate‐General for Employment, Social Affairs and Inclusion.
References
166
Brewer, Mike, and Cormac O’Dea (2012). Measuring living standards with income and consumption: evidence from the UK. ISER Working Paper 2012‐05. Colchester: Institute for Social and Economic Research. Available from https://www.iser. essex.ac.uk/research/publications/working‐papers/iser/2012‐05.pdf.
Brucker, D. L., Mitra, S., Chaitoo, N., Mauro, J. (2015). “More Likely to Be Poor Whatever the Measure: Working‐Age Persons with Disabilities in the United States.” Social Science Quarterly, 96(1): 273‐296.
Burniaux, Jean‐Marc, Thai‐Thahn Dang, Douglas Fore, Michael F. Förster, Marco Mira D’Ercole, and Howard Oxley (1998). Income Distribution and Poverty in Selected OECD countries. OECD Economics Department Working Paper, No. 189. Paris: OECD Publishing. Available from http://www.oecd‐ilibrary.org/docserver/download/ 730801800603.pdf?expires=1483627002&id=id&accname=guest&checksum=5CC73C2BD4284630B03EE3949F9613AF.
Cancho, Cesar, and João Pedro Azevedo (2016). Rent imputation for welfare aggregates: a methodological framework. Presentation at the UNECE seminar on poverty measurement, Geneva, 12‐13 July. Avaiable from http://www.unece.org/index.php?id=41290#.
Castleman, Tony, James Foster, and Stephen S. Smith (2015). Person Equivalent Headcount Measures of Poverty. Institute for International Economic Policy Working Paper Series No. IIEP‐WP‐2015‐10. Washington, D.C.: The George Washington University. Available from https://www2.gwu.edu/~iiep/assets/docs/papers/2015WP/CastlemanFoster SmithIIEPWP201510.pdf.
Chen, Shaohua, and Martin Ravallion (2001). How Did the World's Poorest Fare in the 1990s? Review of Income and Wealth, vol. 47, No. 3, pp. 283‐300.
Citro, Constance F., and Robert T. Michael, eds. (1995). Measuring poverty: A new approach. Washington, D.C.: National Academy Press.
Consolini, Paolo and Donatiello, Gabriella (2013). Improvements of data quality through the combined use of survey and administrative sources and micro simulation model. In The use of registers in the context of EU–SILC: challenges and opportunities. Eurostat methodologies and working paper, Markus Jäntti, Velli‐Matti Törmälehto and Eric Marlier, eds. Luxembourg: Publications Office of the European Union. Available from http://ec.europa.eu/eurostat/documents/3888793/5856365/KS‐TC‐13‐004‐EN.PDF/ fe05381c‐8a77‐4aa9‐ac23‐6c207a3c7f07?version=1.0.
Council of Europe (2005). Concerted development of social cohesion indicators: Methodological guide. Belgium: Council of Europe Publishing.
— (2008). Well‐being for all: Concepts and tools for social cohesion. Trends in Social Cohesion, No. 20. Belgium: Council of Europe Publishing.
Daly, Mary (1992). Europe's poor women? Gender in research on poverty. European Sociological Review, vol. 8, No. 1, pp. 1‐12.
De Vitiis, Claudia, Stefano Falorsi, Francesca Inglese, Alessandra Masi, Nociletta Pannuzi, and Monica Russo (2014). A methodological approach based on indirect sampling to survey the homeless population. Rivista di Statistica Ufficiale, No. 1‐2, pp. 9‐30.
Deaton, Angus (1997). The analysis of household surveys : a microeconometric approach to development policy. Washington, D.C.: The World Bank. Available from http:// documents.worldbank.org/curated/pt/593871468777303124/pdf/multi‐page.pdf.
References
167
Deaton, Angus, and Margaret Grosh (2000). Consumption. In Designing household survey questionnaires for developing countries: lessons from 15 years of the Living Standards Measurement Study, Margaret Grosh and Paul Glewwe, eds. Washington DC: The World Bank. Available from http://documents.worldbank.org/curated/en/ 452741468778781879/pdf/multi‐page.pdf.
Deaton, Angus, and Olivier Dupriez (2011). Purchasing Power Parity Exchange Rates for the Global Poor. American Economic Journal: Applied Economics, vol. 3, No. 2 (April), pp. 137‐166.
Deaton, Angus, and Salman Zaidi (2002). Guidelines for Constructing Consumption Aggregates for Welfare Analysis. LSMS Working Paper 135. Washington, D.C: The World Bank. Available from https://openknowledge.worldbank.org/handle/10986/ 14101.
Delle Fratte, Clodia, and Francesca Lariccia (2015). The impact of Administrative data on final estimates of It‐Silc income variables. Paper presented at the EU‐SILC Best Practice Workshop, 16‐17 September, 2015. London, UK.
Delors, Jacques. (1971). Les indicateurs sociaux. Paris: Futuribles.
Department for Work and Pensions (2004). Review of Households Below Average Income and the Pensioners’ Incomes Series. National Statistics Quality Review Series Report no. 28. London: DWP.
— (2016). Households below average income, 2014/15. London: Department for Work and Pensions. Available from https://www.gov.uk/government/statistics/ households‐below‐average‐income‐199495‐to‐201415.
Dickerson, Andy, and Gurleen Popli (2014). Persistent Poverty and Children's Cognitive Development: Evidence from the UK Millennium Cohort Study. Sheffield Economic Research Paper Series No. 2011023. Sheffield: University of Sheffield. Available from http://www.sheffield.ac.uk/polopoly_fs/1.359958!/file/serps_2011023.pdf.
Economic and Social Council of the United Nations General Assembly (2015). Mainstreaming of the three dimensions of sustainable development throughout the United Nations system. Session July. A/70/75‐E2015/55.
Economic Commission for Latin America and the Caribbean (ECLAC) (2014). Social Panorama 2014. Santiago
European Commission (2010). EUROPE 2020: A Strategy for Smart, Sustainable and Inclusive Growth. Brussels.
— (2012). Employments and social developments in Europe 2012. Luxembourg: Publications Office of the European Union.
European Union Agency for Fundamental Rights (2014). Roma pilot survey technical report: methodology, sampling and fieldwork. Luxembourg: Publications Office of the European Union.
Eurostat (2002). Deuxième Rapport sur le revenu, la pauvreté et l'exclusion sociale. Statistiques Sociales Européennes. Luxembourg : Office des publications officielles des Communautés européennes.
References
168
— (2015a). Smarter, greener, more inclusive? Indicators to support the Europe 2020 strategy. Eurostat Statistical Books, 2015 edition. Luxembourg: Publications Office of the European Union. Available from http://ec.europa.eu/eurostat/ documents/3217494/6655013/KS‐EZ‐14‐001‐EN‐N.pdf/a5452f6e‐8190‐4f30‐8996‐41b1306f7367.
— (2015b). Urban‐rural typology. Statistics Explained. Luxembourg: European Commission. http://ec.europa.eu/eurostat/statistics‐explained/index.php/Urban‐rural_typology.
— (2016). The share of persons at risk of poverty or social exclusion in the EU back to its pre‐crisis level. Eurostat newsrelease 199/2016, 17 October. Luxembourg: European Commission. Available from http://ec.europa.eu/eurostat/documents /2995521/7695750/3‐17102016‐BP‐EN.pdf/30c2ca2d‐f7eb‐4849‐b1e1‐329f48338dc.
Feres, Juan Carlos, and Xavier Mancero (2001). El método de las necesidades básicas insatisfechas (NBI) y sus aplicaciones a América Latina. Series Estudios Estadísticos y Prospectivos, Vol. 4. CEPAL, Naciones Unidas.
Förster, Michael F. (1994). Measurement of Low Incomes and Poverty in A Perspective of International Comparisons. OECD Labour Market and Social Policy Occasional Papers No. 14. Paris: OECD Publishing. Available from http://www.oecd‐ilibrary.org/docserver/download/112854878327.pdf?expires=1483608882&id=id&accname=guest&checksum=871F19D5DF253B1219A98B2DBC9F8CE8.
Foster, James E., Joel Greer and Erik Thorbecke (1984). A class of decomposable poverty measures. Econometrica, vol 52, pp. 761–776.
Fouarge, Didier, and Richard Layte (2005). Welfare Regimes and Poverty Dynamics: The Duration and Recurrence of Poverty Spells in Europe. Journal of Social Policy, Vol. 34, No. 3, pp. 407–426.
Friedman, Milton (1957). A theory of the consumption function. Princeton, NJ: Princeton University Press.
Garner, Thesia I. (2003). A comparison of economic well‐being measures using SIPP and CE data: personal assessments of expenditures and income. 2003 Proceedings of the American Statistical Association, Government Statistics [CD‐ROM], Alexandria, VA: American Statistical Association. Available from https://www.bls.gov/cex/research_papers/pdf/st030130.pdf
Garner, Thesia I., and Randal Verbrugge (2009). Reconciling user costs and rental equivalence: Evidence from US consumer expenditure survey. Journal of Housing Economics, Vol. 18, No. 3 (September), pp. 172‐192.
Garner, Thesia I., and Kathleen Short (2009). Accounting for owner‐occupied dwelling services: Aggregates and distributions. Journal of Housing Economics, Vol. 18, No. 3 (September), pp. 233‐248.
Garner, Thesia I., and Marisa Gudrais (2017). Alternative Poverty Measurement for the U.S: Focus on the Supplemental Poverty Measure Thresholds. Paper presented at the Western Economic Association 13th International Conference. Santiago, January.
Gibson, John (2015). Poverty measurement: We Know Less than Policy Makers Realize. Working Paper in Economics 8/15. Hamilton: University of Waikato. Available from ftp://wms‐webprod1.mngt.waikato. ac.nz/RePEc/wai/econwp/1508.pdf.
References
169
Gibson, John (2005). Statistical Tools and Estimation Methods for Poverty Measures Based on Cross‐Sectional Household Surveys. Chapter 5 in Handbook on Poverty Statistics: Concepts, Methods and Policy Use. New York: United Nations Statistics Division.
Guagnano, Giuseppina, Elisabetta Santerlli, and Isabella Santini (2013). Subjective poverty in Europe: the role of household socioeconomic characteristics and social capital.. Università degli Studi di Roma “La Sapienza” Working paper, No. 113. Rome: Sapienza University. Available from https://core.ac.uk/download/pdf/11978439.pdf.
Guio, Anne‐Catherine, and Eric Marlier (2013). Alternative vs. Current Measure of Material Deprivation at EU Level: What Difference Does it Make? ImPRovE Discussion Paper, No. 13/07. Antwerp: Centre for Social Policy (Antwerp University). Available from http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2616184.
Guio, Anne‐Catherine, Alessio Fusco, and Eric Marlier (2009). A European Union Approach to Material Deprivation using EU‐SILC and Eurobarometer data, IRISS Working Paper 2009‐19, CEPS/INSTEAD, Differdange, Luxembourg. Available from http://iriss.ceps.lu/documents/irisswp117.pdf.
Guio, Anne‐Catherine, David Gordon, and Eric Marlier (2012). Measuring material deprivation in the EU: Indicators for the whole population and child‐specific indicators, Eurostat Methodologies and working papers, 2012 edition. Luxembourg: Office for Official Publications of the European Communities(OPOCE). Available from http://ec.europa.eu/eurostat/documents/3888793/5853037/KS‐RA‐12‐018‐EN.PDF.
Hagenaars, Aldi, and Klaas de Vos (1988). The definition and measurement of poverty. Journal of Human Resources, vol. 23, No. 2 (Spring), pp. 211‐221.
Index of Multiple Deprivations (2007). Measure of multiple deprivations at small area level made up of seven domains. Communities and local Government (GLG). UK: Office for National Statistics.
International Labour Organization (1993). Revision of the international classification of status in employment. 15th International Conference of Labour Statisticians, Geneva, 19‐28 January 1993. Geneva: International Labour Organization. Available from http://www.ilo.org/public/libdoc/ilo/1992/92B09_295_engl.pdf.
— (2004). Resolution Concerning Household Income and Expenditure Statistics, adopted at the 17th International Conference of Labour Statisticians, Geneva, 24 November – 3 December 2004. Geneva: International Labour Organization.
Istat (2011). Services to homeless people. Available from http: //www.istat.it/en/archive/ 45837. Rome: National Institute for Statistics (Istat).
— (2013). The homeless. Available from http://www.istat.it/en/archive/92503. Rome: National Institute for Statistics (Istat).
— (2014). The national research on homeless people conditions. Available from http: //www.istat.it/it/archivio/127256. Rome: National Institute for Statistics (Istat).
Ivanov, Andrey, and Justin Kagin (2014). Roma poverty from a human development perspective. Roma Inclusion Working Papers. Istanbul: UNDP Regional Support Centre for Europe and CIS.
Ivanov, Andrey, Jaroslav Kling and Justin Kagin (2012). Integrated household surveys among Roma populations: one possible approach to sampling used in the UNDP. World Bank‐EC Regional Roma Survey 2011. Roma Inclusion Working Papers. Bratislava: United Nations Development Programme.
References
170
Jäntti, Markus, Veli‐Matti Törmälehto and Eric Marlier (2013). The us of registers in the context of EU‐SILC: challenges and opportunities, Eurostat Statistical Working Papers, 2013 edition. Luxembourg: Publications Office of the European Union. Available from http://ec.europa.eu/eurostat/documents/3888793/5856365/KS‐TC‐13‐004‐EN.PDF/fe05381c‐8a77‐4aa9‐ac23‐6c207a3c7f07?version=1.0.
Jenkins, Stephen P. (1991). Poverty Measurement and the Within‐Household Distribution: Agenda for Action. Journal of Social Policy, vol. 20, No. 4, pp. 457‐483.
Johansson, Sten (1973). The Level of Living Survey: A Presentation. Acta Sociologica, vol. 16, No. 3, pp. 211–219.
Jolliffe, Dean, and Espen Beer Prydz (2016). Estimating international poverty lines from comparable national thresholds. Journal of Economic Inequality, vol. 14, No. 2, pp. 185‐198.
Khan, Seema, Emilie Combaz, and Erika McAslan Fraser (2015). Social exclusion: topic guide. Revised edition. Birmingham, UK: GSDRC, University of Birmingham. Available from http://www.gsdrc.org/topic‐guides/social‐exclusion/dynamics/social‐exclusion‐as‐a‐process.
Klasen, Stephan (2000). Measuring Poverty and Deprivation in South Africa. Review of Income and Wealth, vol. 46, No. 1, pp. 33‐58.
Lanjouw, Peter, Branko Milanovic, and Stefano Paternostro (1998). Poverty and the economic transition: how do changes in economies of scale affect poverty rates for different households. Policy, Research working paper, No. WPS 2009. Washington, DC: World Bank. Available from http://documents.worldbank.org/curated/ en/1998/11/438641/poverty‐economic‐transition‐changes‐economies‐scale‐affect‐poverty‐rates‐different‐households‐vol‐1‐4.
Le Branc, David (2015). Towards integration at last? The sustainable development goals as a network of targets. DESA Working Paper, No. 141 (ST/ESA/2015/141). Available from http://www.un.org/esa/ desa/papers/ 2015/wp141_2015.pdf.
Lenoir, René (1974). Les Exclus: Un Français sur Dix. Paris: Seuil.
Lister, Ruth (1990). The Exclusive Society: Citizenship and the Poor. London: Child Poverty Action Group.
— (2004). Key Concepts: Poverty. Cambridge: Polity Press.
Loewe, Markus, and Nicole Rippin, eds. (2016). The Sustainable Development Goals of the Post 2015 Agenda: Comments on OWG and SDSN proposals. Bonn: German Development Institute.
Ludz, Peter Christian (1971). Materialien zum Bericht zur Lage der Nation. Bonn: Bundesministerium fur innerdeutsche Beziehungen.
MacInnes, Tom, Adam Tinson, Declan Gaffney, Goretti Horgan & Ben Baumberg (2014). Disability, Long Term Conditions and Poverty. New Policy Institute: London. http://www.npi.org.uk/publications/income‐and‐poverty/disability‐long‐term‐conditions‐and‐poverty/
Marlier, Eric, Anthony Atkinson, Bea Cantillon, and Brian Nolan (2007). The EU and social inclusion. Bristol: Policy Press.
References
171
McDonald, Christina, and Katy Negrin (2010). No Data—No Progress: Country Findings. Data Collection in Countries Participating in the Decade of Roma Inclusion. New York: Open Society Foundations. Available from https://www.opensocietyfoundations.org/ sites/default/files/no‐data‐no‐progress‐country‐reports‐20100628_0.pdf.
Meyer, Bruce D., and James X. Sullivan (2011). Further results on measuring the well‐being of the poor using income and consumption, Canadian Journal of Economics, vol. 44, No. 1, pp. 52‐87.
Meyer, Bruce D., Wallace K.C. Mok, and James X. Sullivan (2015). Household surveys in crisis. Journal of Economic Perspectives, vol. 29, No. 4, pp. 199‐226.
Milanovic, Branko (1998). Income, inequality, and poverty during the transition from planned to market economy. World Bank regional and sectoral studies. Washington, D.C. : The World Bank. Available from http://documents.worldbank.org/curated/en/ 229251468767984676/ Income‐inequality‐and‐poverty‐during‐the‐transition‐from‐planned‐to‐market‐economy.
Ministerio de Desarrollo Social ‐ Gobierno de Chile (2015). Nueva Metodología de Medición de Pobreza por Ingresos y Multidimensional, Santiago: Gobierno de Chile, 26 de Enero (Documentos Metodológicos, 28). Available from http://observatorio.ministeriodesarrollosocial.gob.cl/documentos/Nueva_Metodologia_de_Medicion_de_Pobreza.pdf
Morrone, Adolfo, Kate Scrivens, Conal Smith and Carlotta Balestra (2011). Measuring Vulnerability and Resilience in OECD Countries. Paper prepared for the IARIW‐OECD, Conference on Economic Insecurity, Paris France, November 22‐23. Available from http://www.iariw.org/papers/2011/morronepaper.pdf
Narayan, Deepa, Robert Chambers, Meera K. Shah, and Patti Petesch (2000). Voices of the Poor: Crying Out for Change. World Bank Series. Oxford: Oxford University Press.
Nicaise, Ides, and Ingrid Schockaert (2014). The hard to reach among the poor in Europe: lessons from Eurostat's EU‐SILC survey in Belgium. In Hard‐to‐Survey Populations, Roger Tourangeau, Brad Edwards, Tomithy P. Johnson, Kirk M. Wolter and Nancy Bates, eds. Cambridge: Cambridge University Press.
Nolan, Brian and Christopher T. Whelan (2011). Poverty and Deprivation in Europe. Oxford: Oxford University Press.
Noll, Heinz‐Herbert (2007). Household Consumption, Household Incomes and Living Standards – A review of related recent research activities. Mannheim: GESIS – Leibniz Institute for the Social Sciences. Available from https://pdfs.semanticscholar.org/ 4eb1/372c64995fc44520a0145e3fdebd8efd2d09.pdf.
Noll, Heinz‐Herbert, and Weick, Stefan (2014). Major disparities in households’ ability to make ends meet. Comparative analyses of the material standard of living in the European Union. Informationsdienst Soziale Indikatoren ‐ Selected English Articles, 3, pp. 1‐5. (The original German version appeared in ISI 42, 2009: 6–10). Available from http://www.gesis.org/en/services/publications/isi.
ONS (2015). Persistent Poverty in the UK and EU, 2008‐2013. Newport: UK Office for National Statistics. Available from https://www.ons.gov.uk/peoplepopulation andcommunity/personalandhouseholdfinances/incomeandwealth/articles/persistentpovertyintheukandeu/2015‐05‐20.
References
172
— (2016). Effects of taxes and benefits on household income, financial year ending 2015. Newport: UK Office for National Statistics. Available from https://www.ons. gov.uk/peoplepopulationandcommunity/personalandhouseholdfinances/incomeandwealth/bulletins/theeffectsoftaxesandbenefitsonhouseholdincome/financialyearending2015.
Organisation for Economic Co‐operation and Development (1996). Report released at the Thirty‐Fourth High‐Level Meeting of the Development Assistance Committee, Paris, 6‐7 May 1996. Paris.
— (2011). OECD regional typology. Paris: OECD Publishing. Available from https://www.oecd.org/gov/regional‐policy/OECD_regional_typology_Nov2012.pdf.
— (2013a). OECD framework for statistics on the distribution of household income, consumption and wealth. Paris: OECD Publishing. Available from http://www.oecd.org/statistics/302013041e.pdf.
— (2013b). OECD Guidelines for Micro Statistics on Household Wealth. Paris: OECD Publishing. Available from http://www.oecd.org/statistics/OECD‐Guidelines‐for‐Micro‐Statistics‐on‐Household‐Wealth.pdf.
— (2013c). OECD Guidelines on Measuring Subjective Well‐being. Paris: OECD Publishing. Available from http://dx.doi.org/10.1787/9789264191655‐en.
— (2014). Society at a Glance 2014: OECD Social Indicators. Paris, France: OECD Publishing.
— (2014). National Accounts at a Glance 2014, OECD Publishing. Available from http://dx.doi.org/10.1787/na_glance‐2014‐en
Orshansky, Mollie (1963). Children of the Poor. Social Security Bulletin, vol. 26, No. 7, pp. 3‐13.
Pakistani Ministry of Planning, Development, and Reform (2016). Multidimensional Poverty in Pakistan, Official Report. Islamabad/Oxford: Pakistani Ministry of Planning, Development, and Reform/ United Nations Development Programme Pakistan/ Oxford Poverty & Human Development Initiative.
Ponthieux, Sophie (2013). Income pooling and equal sharing within the household — What can we learn from the 2010 EU‐SILC module? Eurostat Methodologies and Working paper, 2013 edition. Luxembourg: Eurostat. Available from http://ec.europa.eu/ eurostat/documents/3888793/5857781/KS‐RA‐13‐013‐EN.PDF.
Ravallion, Martin, and Shaohua Chen (2009). Weakly relative poverty. Policy Research Working Paper, No. 4844. Washington DC: The World Bank. Available from https://openknowledge.worldbank.org/bitstream/handle/10986/4168/WPS4844.pdf?sequence=1.
Ravallion, Martin, Shaohua Chen, and Prem Sangraula (2009). Dollar a day revisited. World Bank Economic Review, vol. 23, No. 2, pp. 163‐184.
Ravallion, Martin, Gaurav Datt, and Dominique van de Walle (1991). Quantifying absolute poverty in the developing world. Review of Income and Wealth, vol. 37, No. 4, pp. 345‐361.
Ravallion, Martin (2012). Poor, or just feeling poor? On using subjective data in measuring poverty. World Bank Policy Research Working Paper, No. 5968. Washington, D.C: The World Bank. Available from http://documents.worldbank.org/curated/en/ 560421468152099804/pdf/WPS5968.pdf.
References
173
— (2015). Toward Better Global Poverty Measures. CGD Working Paper, No. 417. Washington, DC: Center for Global Development. Available from http://www. cgdev.org/publication/toward‐better‐global‐poverty‐measures‐working paper‐417.
Renwick, Trudi, Bettina Aten, Eric Figueroa and Troy Martin (2014). Supplemental Poverty Measure: a comparison of geographic adjustments with regional price parities vs. median rents from the American Community Survey. SEHSD Working Paper No. 2014‐22. United States Census Bureau. Available from https://www.census.gov/hhes/povmeas/methodology/supplemental/research/SPMUsingRPPs.pdf
Renwick, Trudi and Liana Fox (2016). The Supplemental Poverty Measure: 2015. Current Population Reports. United States Census Bureau. Available from https://www.census.gov/content/dam/Census/library/publications/2016/demo/p60‐258.pdf
Rio Group (2006). Compendium of best practices in poverty measurement. Rio de Janeiro: Brazilian Institute for Geography and Statistics. ISBN 85‐240‐3908‐6. Available from http://www.ibge.gov.br/poverty/.
Robine, Jean‐Marie, Emmanuelle Cambois, Wilma Nusselder, Bernard Jeune, Herman Van Oyen, and Carol Jagger (2013). The joint action on healthy life years (JA: EHLEIS). Archives of Public Health, Vol. 71, No. 2.
Rowntree, B. Seebohm (1901). Poverty: A Study of Town Life. London: Macmillan and Co.
Santos, Maria Emma, Pablo Villatoro, Xavier Mancero, and Pascual Gerstenfeld (2015). A Multidimensional Poverty Index for Latin America. OPHI Working Paper, No 79. OPHI: Oxford University. Available from https://www.ophi.org.uk/wp‐content/uploads/OPHIWP079.pdf.
Save the Children (2010). A Fair Chance At Life: Why Equity Matters for Child Mortality. Save the Children: London, UK.
Sen, Amartya K (1983). Choice, Welfare, and Measurement. Oxford: Basil Blackwell.
— (2000). Social Exclusion: Concept, Application and Scrutiny. Social Development Papers No.1. Office of Environment and Social Development, Asian Development Bank. Available from https://www.adb.org/sites/default/files/ publication/29778/social‐exclusion.pdf.
— (1997). From Income Inequality to Economic Inequality. Southern Economic Journal, vol. 64, No. 2, pp. 383‐401.
Serafino, Paola, and Richard P. Tonkin (2017a). Comparing poverty estimates using income, expenditure and material deprivation. In Monitoring Social Europe, Anthony B. Atkinson, Anne‐Catherine Guio and Eric Marlier, eds. Luxembourg: Eurostat.
— (2017b). Statistical matching of EU‐SILC and the Household Budget Survey to compare poverty estimates using income, expenditure and material deprivation, Eurostat methodologies and working paper, 2016 edition. Luxembourg: Eurostat.
Silver, Hilary (1995). Reconceptualizing Social Disadvantage: Three paradigms of social exclusion. In Social Exclusion: Rhetoric, reality, responses, Gerry Rogers, Charles Gore and Jose B. Figueiredo, eds. Geneva: International Institute for Labour Studies. United Nations Development Programme.
References
174
Škobla, Daniel, Tadas Leončikas, and Martina Štěpánková (2009). Ethnicity as a statistical indicator for the monitoring of living conditions and discrimination. Analytical report and recommendationsfor the Slovak Republic. Valeur: United Nations Development ProgrammeRegional Bureau for Europeand the Commonwealth of Independent States.
Smith, Adam (1812). An Inquiry into the Natural Causes of the Wealth of Nations. London: Ward, Lock and Co. (first published 1776).
Smith, Tom, Michael Noble, Stefan Noble, Gemma Wright, David McLennan, and Emma Plunkett (2015a). The English Indices of Deprivation 2015; Technical Report. Accessed at: https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/464485/English_Indices_of_Deprivation_2015_‐_Technical‐Report.pdf.
— (2015b). The English Indices of Deprivation 2015; Research Report. London: Department for Communities and Local Government. Available from https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/464597/English_Indices_of_Deprivation_2015_‐_Research_Report.pdf.
Stewart, Frances (1985). Basic Needs in Developing Countries. Baltimore: Johns Hopkins University Press.
Stewart, Frances, Manuel Barron, Graham Brown, and Marcia Hartwell (2006). Social exclusion and conflict: analysis and policy implications. CRISE Policy Paper, Queen Elizabeth House, Oxford University, January 2005 (updated April 2006).
Stiglitz, Joseph E., Amartya Sen, and Jean‐Paul Fitoussi (2009). Report by the Commission on the Measurement of Economic Performance and Social Progress. Available from http://www.stiglitz‐sen‐fitoussi.fr/en/index.htm.
Streeten, Paul, Shahid Javed Burki, Mahbub Ul Haq, Norman Hicks, and Frances Stewart (1981). First things first: Meeting basic human Needs in Developing Countries. Oxford: Oxford University Press.
Suppa, Nicolai (2015). Towards a Multidimensional Poverty Index in Germany. OPHI Working Paper No. 98. OPHI: Oxford University.
Tim Goedemé, Bérénice Storms, Tess Penne & Karel Van Den Bosch (2015). Pilot project for the development of a common methodology on reference budgets in Europe. European Commission: Brussels. Available from http://ec.europa.eu/social/BlobServlet?docId=14917&langId=en
Tonkin, Richard P., Veli‐Matti Törmälehto, Nathan Thomas, James Lewis, and Mira Kajantie (2014). Social Transfers in Kind in the United Kingdom and Finland: Micro‐level Measurement and Distributional Impact. Paper presented at the IARIW 33rd General Conference, 24‐30 August 2014. Rotterdam, the Netherlands. Available from http://www.iariw.org/papers/2014/tonkinPaper.pdf.
Tonkin, Richard P., Paola Serafino, and Ruth E. Davies (2016). Using statistical matching to facilitate the comparison of poverty estimates using income, consumption and wealth. Paper presented at the UNECE seminar on poverty measurement, Geneva, 12‐13 July 2016. Available from http://www.unece.org/fileadmin/DAM/stats/ documents/ece/ ces/ge.15/2016/Sem/WP7_UK_ENG.pdf.
References
175
Törmälehto, Veli‐Matti, and Hannele Sauli (2013). The distributional impact of imputed rent in EUSILC 2007‐2010. Eurostat methodologies and working paper, 2013 edition. Luxembourg: Eurostat. Available from http://ec.europa.eu/eurostat/documents/ 3888793/5857525/KS‐RA‐13‐011‐EN.PDF/516f90a7‐25d4‐4359‐a297‐fa8311758a1a.
Törmälehto, Veli‐Matti, Olli Kannas and Markku Säylä (2013). Integrated measurement of household‐level income, wealth and non‐monetary well‐being in Finland. Statistics Finland working paper. Available from http://www.stat.fi/tup/julkaisut/tiedostot/julkaisuluettelo/ywrp1_201300_2013_10518_net.pdf
Tourangeau, Roger, Brad Edwards, Tomithy P. Johnson, Kirk M. Wolter, and Nancy Bates, eds. (2014). Hard‐to‐Survey Populations. Cambridge: Cambridge University Press.
Townsend, Peter (1979). Poverty in the United Kingdom: A survey of household resources and standards of living. London: Allen Lane and Penguin Books.
United Nations (2008). Assessment of capacity for countries of Eastern and South‐Eastern Europe, Caucasus and Central Asia to produce MDG‐relevant statistics. Geneva, Switzerland: United Nations.
United Nations Development Programme (1991). Human Development Report. New York/Oxford: United Nations Development Programme.
1. (2010). What will it take to achieve the millennium development goals? An international assessment. New York: United Nations Development Programme.
2. (2010). The real wealth of nations: Pathways to human development. Human Development Report 2010. New York: United Nations Development Programme.
3. (2012). Human Development Report of the Kyrgyz Republic (2012). Bishkek: United Nations Development Programme in the Kyrgyz Republic.
United Nations Development Programme and Eurasian Development Bank (2015). Labour Migration, Remittances and Human Development in Central Asia. Central Asia Human Development Series.
United Nations Development Programme Istanbul Regional Hub (2016). Meeting report of the Roundtable Discussion on the 2030 Agenda: Moving Forward in Mainstreaming, Implementation, Monitoring and Reporting of the SDGs, Istanbul, 25‐27 April 2016. Istanbul.
United Nations Development Programme Regional Bureau for Europe and the Commonwealth of Independent States (2011). Beyond Transition: Towards Inclusive Societies. Europe and Central Asia Human Development Report. Bratislava. Available from http://hdr.undp.org/sites/default/ files/rbec_hdr_2011_en.pdf.
United Nations Economic and Social Council (1998). Statement of commitment for action to eradicate poverty adopted by Administrative Committee on Coordination. Press release ECOSOC/5759, May 20. Accessed December 20, 2016. Available from http://www.un.org/press/en/1998/19980520.eco5759.html.
United Nations Economic Commission for Europe (2006). Conference of European Statisticians. Recommendations for the 2010 Censuses of Population and Housing. New York and Geneva: United Nations. ECE/CES/STAT/NONE/2006/4. Available from https://www.unece.org/fileadmin/DAM/stats/publications/CES_2010_Census_Recommendations_English.pdf.
References
176
— (2011). Canberra Group Handbook on Household Income Statistics: Second Edition. Geneva: United Nations.
— (2012a). Conference of European Statisticians. In‐depth review of poverty statistics, Note prepared by the State Statistics Service of Ukraine and Eurostat, Geneva, 5‐6 November 2012. Geneva. ECE/CES/BUR/2012/NOV/4.
— (2012b). Outcome of the in‐depth review of poverty statistics. Geneva. Available from www.unece.org/fileadmin/DAM/ stats/documents/ece/ces/2013/7.pdf.
— (2012c). Survey Module for Measuring Health State. United Nations: New York and Geneva. Available from http://www.unece.org/index.php?id=31580.
— (2013). Getting the facts right: A guide to presenting metadata with examples on millennium development goal indicators. ECE/CES/29. New York and Geneva: United Nations.
— (2015a). Conference of European Statisticians Recommendations for the 2020 Censuses of Population and Housing. New York and Geneva: United Nations. ECE/CES/41. Available from https://www.unece.org/ fileadmin/DAM/stats/publications/2015/ ECE_CES_41_WEB.pdf.
— (2015b). Indictors of gender equality. ECE/CES/37. New York and Geneva: United Nations.
United Nations Environment Programme/ United Nations Conference on Trade and Development (1974). The Cocoyoc Declaration. Adopted by the Participants in the UNEP/UNCTAD Symposium on ‘Patterns of Resource Use, Environment and Development Strategies’. International Organization, Vol. 29, No. 3, pp. 893–901.
United Nations Children's Fund (2010). Serbia Multiple Indicator Cluster Survey 2010. Belgrade: Statistical Office of the Republic of Serbia and UNICEF.
— (2011). Republic of Macedonia Multiple Indicator Cluster Survey 2011. Skopje: Ministry of Health/Ministry of Education and Science/Ministry of Labour and Social Policy.
— (2012). Bosnia and Herzegovina Multiple Indicator Cluster Survey 2011‐2012. Final Report. Sarajevo: UNICEF Office for Bosnia and Herzegovina/ The Agency for Statistics of Bosnia and Herzegovina, the Federal Ministry of Health, the Ministry of Health and Social Welfare of the Republic of Srpska and the Institute for Public Health of the Federation of Bosnia and Herzegovina.
— (2014a). Montenegro and Montenegro Roma Settlements Multiple Indicator Cluster Survey 2013. Podgorica: Statistical Office of Montenegro (MONSTAT) and UNICEF.
— (2014b). Serbia and Serbia Roma Settlements Multiple Indicator Cluster Survey and Multiple Indicator Cluster Survey 2014. Belgrade: Statistical Office of the Republic of Serbia and UNICEF.
United Nations Statistical Commission (2016a). Report of the Inter‐agency and Expert Group on Sustainable Development Goal Indicators, New York, 8‐11 March 2016 (Revised). New York. E/CN.3/2016/2/Rev.1.
— (2016b). Report of the forty‐seventh session, New York, 8‐11 March 2016. New York. E/2016/24E/CN.3/2016 /34.
Vaz, Ana (2014). Are children among the poorest? OPHI Policy Briefing. Oxford: OPHI. Available from http://www.ophi.org.uk/wp‐content/uploads/Age‐decomp‐briefing.pdf
References
177
Webber, Dominic, and Richard P. Tonkin (2013). Statistical matching of EU‐SILC and the Household Budget Survey to compare poverty estimates using income, expenditure and material deprivation, Eurostat methodologies and working paper, 2013 edition. Luxembourg: Eurostat. Available from http://ec.europa.eu/eurostat/ documents/3888793/5857145/KS‐RA‐13‐007‐EN.PDF/37d4ffcc‐e9fc‐42bc‐8d4f‐fc89c65ff6b1.
Wolff, Johatan, and Avner De‐Shalit (2007). Disadvantage. Oxford: Oxford University Press.
World Bank (1990). World Development Report 1990: Poverty. New York: Oxford University Press. Available from https://openknowledge.worldbank.org/bitstream/handle/ 10986/5973/WDR%201990%20‐%20English.pdf?sequence=5&isAllowed=y.
— (2000). Making transition work for everyone: poverty and inequality in Europe and Central Asia. A World Free of Poverty series. Washington DC: World Bank. Available from http://documents.worldbank.org/curated/en/571801468037741599/ Making‐transition‐work‐for‐everyone‐poverty‐and‐inequality‐in‐Europe‐and‐Central‐Asia.
— (2017). Monitoring global poverty. Report of the Commission on Global Poverty. Washington DC: World Bank. Available from http://www.worldbank.org/en/programs/commission‐on‐global‐poverty.
Zacharias, Ajit, Thomas Masterson, and Emil Memis (2014). Time Deficits and Poverty. The Levy Institute Measure of Time and Consumption Poverty for Turkey. United Nations Development Programme. Available form http://www.levyinstitute.org/pubs/rpr_5 _14.pdf.
178
AnnexI Goal1andGoal10poverty‐relatedtargetsandindicatorsinthe2030AgendaforSustainableDevelopment
Goals and Targets Global Indicators Tier Possible Custodian Agency(ies)
Goal 1. End poverty in all its forms everywhere 1.1 By 2030, eradicate extreme poverty for all people everywhere, currently measured as people living on less than $1.25 a day
1.1.1 Proportion of population below the international poverty line, by sex, age, employment status and geographical location (urban/rural)
1 World Bank
1.2 By 2030, reduce at least by half the proportion of men, women and children of all ages living in poverty in all its dimensions according to national definitions
1.2.1 Proportion of population living below the national poverty line, by sex and age
1 National Governments
1.2.2 Proportion of men, women and children of all ages living in poverty in all its dimensions according to national definitions
2 National Governments
1.3 Implement nationally appropriate social protection systems and measures for all, including floors, and by 2030 achieve substantial coverage of the poor and the vulnerable
1.3.1 Proportion of population covered by social protection floors/systems, by sex, distinguishing children, unemployed persons, older persons, persons with disabilities, pregnant women, newborns, work injury victims and the poor and the vulnerable
1 ILO
1.4 By 2030, ensure that all men and women, in particular the poor and the vulnerable, have equal rights to economic resources, as well as access to basic services, ownership and control over land and other forms of property, inheritance, natural resources, appropriate new technology and financial services, including microfinance
1.4.1 Proportion of population living in households with access to basic services
3
AnnexI
179
Goals and Targets Global Indicators Tier Possible Custodian Agency(ies)
1.4.2 Proportion of total adult population with secure tenure rights to land, with legally recognized documentation and who perceive their rights to land as secure, by sex and by type of tenure
3 World Bank as part of 23 members of Global Donor Working Group on Land
1.5 By 2030, build the resilience of the poor and those in vulnerable situations and reduce their exposure and vulnerability to climate‐related extreme events and other economic, social and environmental shocks and disasters
1.5.1 Number of deaths, missing persons and persons affected by disaster per 100,000 people a
2 UNISDR
1.5.2 Direct disaster economic loss in relation to global gross domestic product (GDP)
2 UNISDR
1.5.3 Number of countries with national and local disaster risk reduction strategies
2 UNISDR
1.a Ensure significant mobilization of resources from a variety of sources, including through enhanced development cooperation, in order to provide adequate and predictable means for developing countries, in particular least developed countries, to implement programmes and policies to end poverty in all its dimensions
1.a.1 Proportion of resources allocated by the government directly to poverty reduction programmes
2 World Bank
1.a.2 Proportion of total government spending on essential services (education, health and social protection)
3 World Bank
1.b Create sound policy frameworks at the national, regional and international levels, based on pro‐poor and gender‐sensitive development strategies, to support accelerated investment in poverty eradication actions
1.b.1 Proportion of government recurrent and capital spending to sectors that disproportionately benefit women, the poor and vulnerable groups
3
Goal 10. Reduce inequality within and among countries 10.1 By 2030, progressively achieve and sustain income growth of the bottom 40 per cent of the population at a rate higher than the national average
10.1.1 Growth rates of household expenditure or income per capita among the bottom 40 per cent of the population and the total population
1 World Bank
AnnexI
180
Goals and Targets Global Indicators Tier Possible Custodian Agency(ies)
10.2 By 2030, empower and promote the social, economic and political inclusion of all, irrespective of age, sex, disability, race, ethnicity, origin, religion or economic or other status
10.2.1 Proportion of people living below 50 per cent of median income, by age, sex and persons with disabilities
3 World Bank
10.3 Ensure equal opportunity and reduce inequalities of outcome, including by eliminating discriminatory laws, policies and practices and promoting appropriate legislation, policies and action in this regard
10.3.1 Proportion of the population reporting having personally felt discriminated against or harassed within the previous 12 months on the basis of a ground of discrimination prohibited under international human rights law
3 OHCHR
Source: UNSTAT. Report of the Inter‐agency and Expert Group on Sustainable Development Goal Indicators (Revised) E/CN.3/2016/2/Rev.1. http://bit.ly/1N9Ynvg, IAEG. Provisional proposed tiering system for the indicators, as of March 24, 2016. http://bit.ly/1qjJdcD
Notes: Tiers of indicators reflect their conceptual clarity and availability, in particular: Tier 1: Indicator conceptually clear, established methodology and standards available and data regularly produced by countries. Tier 2: Indicator conceptually clear, established methodology and standards available but data are not regularly produced by countries. Tier 3: Indicator for which there are no established methodology and standards or methodology/standards are being developed/tested.
181
AnnexII ResultsoftheUNECEsurveyonpovertymeasurement
In 2014, UNECE conducted a survey on poverty measurement among national statistical offices of countries participating in the work of the Conference of European Statisticians (these include UNECE member countries, see http://www.unece.org/oes/nutshell/member_states_representatives.html and additionally Australia, Brazil, Chile, China, Colombia, Japan, Mexico, Mongolia, New Zealand and Republic of Korea). The survey inquired about practices, standards and techniques in poverty measurement. The tables below display the results from the 45 countries who responded. Table II.1 Absolute poverty
Country Poverty indicator
Welfare indicator
Poverty threshold (line)
Coverage Unit of analysis
Use of equivalence scale
Available breakdowns
Possibility of measurement of persistent poverty
Periodicity of indicator published
Type of data source
European Statistical System
Estonia Income/ disposable income
Monetary and in kind
Subsistence minimum
Private households
Person OECD scale Demographic, geographic, social
Yes Annual Sample survey
Hungary Income/ disposable income
Monetary and in kind
Subsistence minimum
Private households
Persons Special Hungarian scale1
Demographic Yes Annual Sample survey
Italy Expenditure/ consumption expenditure
Monetary and in kind
Cost of basic needs
Private households
Household, person
Implicit equivalence scale
Demographic, geographic, social
No Annual Sample survey
Latvia Income/ disposable income
Monetary Subsistence minimum
2
Private households
Person No Only country level
No Monthly Sample survey
Netherlands14 Income/ disposable income
Monetary and in kind
Fixed amount (990 euro a month in prices of 2012)
Private households3
Household, person
National equivalence scale
Demographic, geographic, social
Yes Annual Sample survey
AnnexII
182
Country Poverty indicator
Welfare indicator
Poverty threshold (line)
Coverage Unit of analysis
Use of equivalence scale
Available breakdowns
Possibility of measurement of persistent poverty
Periodicity of indicator published
Type of data source
Income/ disposable income
Monetary and in kind
Minimum benefits policy bound (amounts vary by household type)
Private households
Household, person
National equivalence scale
Demographic, geographic, social
Yes Annual Sample survey
Poland Expenditure/ consumption expenditure
Monetary and in kind
Subsistence minimum
Private households
Person OECD scale Demographic, geographic, social
No Annual Sample survey
Expenditure/ consumption expenditure
Monetary and in kind
”Legal” poverty threshold4
Private households
Person No Demographic, geographic, social
No Annual Sample survey
Switzerland Income/ disposable income
Monetary and in kind
Private households
Person No Demographic, social
Yes Annual Sample survey
United Kingdom14
Income/ disposable income
Monetary and in kind
2010/11, 60% median income before and after housing costs
Private Households
Person Modified OECD scale
Demographic, geographic, social
No Annual Sample survey
Eastern Europe, Caucasus and Central Asia
Armenia Expenditure/ consumption expenditure
Monetary and in kind
Cost of basic needs
Private households
Household, person
National equivalence scale5
Demographic, geographic
No Annual Sample survey
Azerbaijan Expenditure/ consumption expenditure
Monetary Subsistence minimum
Private households
Person No Demographic, geographic
No Annual Sample survey
AnnexII
183
Country Poverty indicator
Welfare indicator
Poverty threshold (line)
Coverage Unit of analysis
Use of equivalence scale
Available breakdowns
Possibility of measurement of persistent poverty
Periodicity of indicator published
Type of data source
Belarus Income/ disposable income
Monetary and in kind
Subsistence minimum
Private households
Person No Demographic, geographic, Social
No Quarter, annual
Sample survey
Kazakhstan Expenditure/ consumption expenditure
Monetary and in kind
Subsistence minimum
Private households
Person National equivalence scale6
Demographic, geographic
No Quarter, annual
Sample survey
Kyrgyzstan Expenditure/ consumption expenditure
Monetary and in kind
Cost of basic needs
Private households
Person No Demographic, geographic
No Annual Sample survey
Republic of Moldova
Expenditure/ consumption expenditure
Monetary and in kind
Cost of basic needs
Private households
Person OECD scale Demographic, geographic
Partially Annual Sample survey
Russian Federation
Income/macroeconomic indicator of per capita disposable income
Monetary Subsistence minimum7
Private households
Person No Quarter ‐ Demographic, geographic; Year ‐ geographic
No Quarter, annual
Sample survey
Income/ disposable Income
Monetary and in kind
Subsistence minimum
Private households
Person No Demographic, geographic, social
No Annual Sample survey
Income/ disposable income
Monetary and in kind
1.25$, 2.5$ and 4.0$ per capita per day according to PPP
Private households
Person No Only country level
No Annual Sample survey
AnnexII
184
Country Poverty indicator
Welfare indicator
Poverty threshold (line)
Coverage Unit of analysis
Use of equivalence scale
Available breakdowns
Possibility of measurement of persistent poverty
Periodicity of indicator published
Type of data source
Ukraine Income/ disposable income
Monetary and in kind
Subsistence minimum
Private households
Household, person
National equivalence scale8
Demographic, geographic, social
Yes Quarter, annual
Sample survey
Expenditure/ consumption expenditure
Monetary and in kind
Subsistence minimum
Private households
Household, person
National equivalence scale
Demographic, geographic, social
Yes Quarter, annual
Sample survey
Expenditure/ consumption expenditure
Monetary and in kind
5.0$ per equivalent person per day according to PPP
Private households
Person National equivalence scale
Only country level
Yes Quarter, annual
Sample survey
Uzbekistan Expenditure/ consumption expenditure
Monetary and in kind
Cost of basic needs
Private households
Person No Demographic, geographic, social
No Annual Sample survey
Other countries
Bosnia and Herzegovina
Expenditure/ consumption expenditure
Monetary and in kind
Cost of basic needs
Private households
Household, person
OECD scale Demographic, geographic, social
No Every 3‐4 years
Sample survey
Canada14 Income/ disposable income
Monetary Income level where households spend 20% more of their income on necessities than average household
Private households
Economic Family9
Thresholds are determined for each family size
Demographic, geographic, social
No Annual Sample survey
AnnexII
185
Country Poverty indicator
Welfare indicator
Poverty threshold (line)
Coverage Unit of analysis
Use of equivalence scale
Available breakdowns
Possibility of measurement of persistent poverty
Periodicity of indicator published
Type of data source
Income/ disposable income
Monetary Cost of basic needs
Private households
Economic Family
Square root scale
Demographic, geographic, social
No Annual Sample survey
China Income/ disposable income
Monetary and in kind
Minimum standard set by the central government
Rural households
Person No Demographic, geographic
Yes Annual Sample survey
Colombia Income/ disposable income
Monetary and in kind
Cost of basic needs
Private households
Person No Demographic, geographic, social
No Annual Sample survey
Mexico Income/ disposable income
Monetary and in kind
Cost of basic needs
Private households
Household, person
Yes Demographic, geographic, social
Yes Every 2 years for federal entities, every 5 years for municipa‐lities
Sample survey
Mongolia Expenditure/ consumption expenditure
Monetary and in kind
Cost of basic needs
Private households
Household, person
No Demographic, geographic, social
No Annual Sample survey
Montenegro Expenditure/ consumption expenditure
Monetary and in kind
Cost of basic needs
Private households
Person Modified OECD scale
Demographic, geographic, social
No Annual Sample survey
AnnexII
186
Country Poverty indicator
Welfare indicator
Poverty threshold (line)
Coverage Unit of analysis
Use of equivalence scale
Available breakdowns
Possibility of measurement of persistent poverty
Periodicity of indicator published
Type of data source
New Zealand14 Income/ disposable income
Monetary 60% of 2007 median household equivalised disposable income
Private households
Person Revised Jensen Scale (1988)10
Demographic No Annual Sample survey
Income/ disposable income
Monetary 60% of 2007 median household equivalised disposable income less 25% as an allowance for housing costs
Private households
Person Revised Jensen Scale (1988)
Demographic No Annual Sample survey
Turkey Expenditure/ consumption expenditure
Monetary and in kind
1$, 2.15$ and 4.30$ per capita per day according to PPP
Private households
Person National equivalence scale11
Geographic No Annual Sample survey
United States Income Monetary Cost of all goods and services (estimated as 3 X the cost of basic food plan) (see Box 3.10)
Private households
Person Implicit in food plans
Demographic, geographic, social
No Annual Sample survey
AnnexII
187
Country Poverty indicator
Welfare indicator
Poverty threshold (line)
Coverage Unit of analysis
Use of equivalence scale
Available breakdowns
Possibility of measurement of persistent poverty
Periodicity of indicator published
Type of data source
Income/ disposable income
Monetary and in kind
30th‐36th percentile of expenditures on food, clothing, shelter and utlities12(see Box 3.10)
Private households
Person National equivalence scale13
Demographic, geographic, social
No Annual Sample survey
Source: National Statistical Offices (2015). Notes: 1 First active adult ‐ 1.00; all other adults ‐ 0.75; first child ‐ 0.65; second child ‐ 0.50; all other children ‐ 0.40 unit. In case of inactive households: first inactive adult ‐ 0.90;
all other persons ‐ 0.65; 2 From 2014 the CSB terminates calculation of subsistence minimum; 3 Do not include student households and households with (annual) incomplete income data; 4 Is the amount which, according to the Law on Social Assistance (uniform text Journal of Laws 2013, item 182, with later amendments), provides eligibility for a monetary benefit from social assistance system.; 5 Children 0‐14 ‐ 0.65, Others ‐ 1; 6 First member of a household ‐ 1, each second and subsequent member of a household ‐ 0.8; 7 ‐ population with incomes below the subsistence minimum – short‐time and preliminary assessment (by the analytical model). The survey (HBS) will be conducting till 2015. 8 First member of a household ‐ 1, each second and subsequent member of a household ‐ 0.7; 9 An economic family is a group of individuals sharing a common dwelling unit who are related by blood, marriage (including common‐law relationships) or adoption; 10 Takes into account number and age of children; 11 (Number of adults + 0.9 * Number of children) ^ 0.6; 12 The SPM differs from the official poverty measure by taking account of government in‐kind benefits and necessary expenses and taxes that are not in the official measure, and also adjusts thresholds geographically and by housing tenure type (i.e. owners with a mortgage, owners without a mortgage, and renters). The SPM thresholds are based on a range in the distribution of the sum of annual expenditures for food, clothing, shelter, and utilities (FCSU), with a multiplier for other basic needs. 13 The three‐parameter scale is calculated in the following way: One and two adults: scale = (adults)0.5 Single parents: scale = (adults + 0.8*first child + 0.5*other children)0.7 All other families: scale = (adults + 0.5*children)0.7; 14 Use anchored poverty lines, which are distinct from absolute poverty lines (see Table 2.1 and Chapter 3).
AnnexII
188
Table II.2 Relative poverty Country Poverty
indicator
Welfare
indicator
Poverty Threshold
(line)
Coverage Unit of
analysis
Use of
equivalence
scale
Available
breakdowns
Possibility of
measurement
of persistent
poverty
Periodicity of
indicator
published
Type of
data
source
European Statistical System
Austria Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Person Modified
OECD scale
Demographic,
social
Yes Annual Sample
survey
Bulgaria Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Household,
person
Modified
OECD scale
Demographic,
geographic,
social
Yes Annual Sample
survey
Croatia Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Household,
person
Modified
OECD scale
Demographic,
geographic,
social
Yes Annual Sample
survey
Cyprus Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Persons Modified
OECD scale
Demographic,
geographic,
social
Yes Annual Sample
survey
Czechia Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Person Modified
OECD scale
Demographic,
social
Yes Annual Sample
survey
Denmark Income/
disposable
income
Monetary 50% and 60% of
the equivalised
median disposable
income
Total
population
Person Modified
OECD scale
Demographic,
social
Yes Annual Sample
survey
AnnexII
189
Country Poverty
indicator
Welfare
indicator
Poverty Threshold
(line)
Coverage Unit of
analysis
Use of
equivalence
scale
Available
breakdowns
Possibility of
measurement
of persistent
poverty
Periodicity of
indicator
published
Type of
data
source
Estonia Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Person Modified
OECD scale
Demographic,
geographic,
social
Yes Annual Sample
survey
Germany Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Person Modified
OECD scale
Demographic,
social
Yes Annual Sample
survey
Hungary Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Persons Modified
OECD scale
Demographic Yes Annual Sample
survey
Ireland Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Persons National
equivalence
scale1
Demographic,
geographic,
social
Yes Annual Sample
survey
Italy Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income.
Private
households
Household,
person
Modified
OECD scale
Demographic,
geographic,
social
Yes Annual Sample
survey
Expenditure/
consumption
expenditure
Monetary
and in kind
60% of the
equivalised
median
consumption
Private
households
Household,
person
Carbonaro’s
equivalence
scale3
Demographic,
geographic,
social
No Annual Sample
survey
Latvia Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Household,
person
Modified
OECD scale
Demographic,
geographic,
social
Yes Annual Sample
survey
AnnexII
190
Country Poverty
indicator
Welfare
indicator
Poverty Threshold
(line)
Coverage Unit of
analysis
Use of
equivalence
scale
Available
breakdowns
Possibility of
measurement
of persistent
poverty
Periodicity of
indicator
published
Type of
data
source
Lithuania Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Person Modified
OECD scale
Demographic,
geographic,
social
Yes Annual Sample
survey
Netherlands Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households4
Household,
person
National
equivalence
scale
Demographic,
geographic,
social
Yes Annual Sample
survey
Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Household,
person
Modified
OECD scale
Demographic,
geographic,
social
Yes Annual Sample
survey
Norway Income/
disposable
income
Monetary 50% and 60% of
the equivalised
median disposable
income
Private
households
Person Modified
OECD scale
Demographic,
geographic,
social
Yes Annual Census,
administr
ative
records
Poland Expenditure/
consumption
expenditure
Monetary
and in kind
50% of the
equivalised mean
consumption
expenditure
Private
households
Person OECD scale Demographic,
geographic,
social
No Annual Sample
survey
Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Person Modified
OECD scale
Demographic,
geographic,
social
Yes Annual Sample
survey
Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Person Modified
OECD scale
Demographic,
geographic,
social
Yes Every 4‐5
years
Sample
survey
AnnexII
191
Country Poverty
indicator
Welfare
indicator
Poverty Threshold
(line)
Coverage Unit of
analysis
Use of
equivalence
scale
Available
breakdowns
Possibility of
measurement
of persistent
poverty
Periodicity of
indicator
published
Type of
data
source
Romania Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Household,
persons
Modified
OECD scale
Demographic,
geographic,
social
Yes Annual Sample
survey
Slovakia Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Household,
persons
Modified
OECD scale
Demographic,
geographic,
social
Yes Annual Sample
survey
Spain Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Person Modified
OECD scale
Demographic,
geographic,
social
Yes Annual Sample
survey
Switzerland Income/
disposable
income
Monetary
and in kind
50% and 60% of
the equivalised
median disposable
income
Private
households
Person Modified
OECD scale
Demographic,
geographic,
social
Yes Annual Sample
survey
United
Kingdom
Income/
disposable
income
Monetary 60% median
income before and
after housing costs
Private
Households
Person Modified
OECD scale
Demographic,
geographic,
social
No Annual Sample
survey
Eastern Europe, Caucasus and Central Asia
Belarus Income/ disposable income
Monetary and in kind
60% of the equivalised median disposable income
Private households
Person National equivalence scale5
Geographic No Annual Sample survey
Georgia Expenditure/ consumption expenditure
Monetary and in kind
60% of equivalised median consumption expenditure
Private households
Person National equivalence scale6
Demographic, geographic
No Annual Sample survey
AnnexII
192
Country Poverty
indicator
Welfare
indicator
Poverty Threshold
(line)
Coverage Unit of
analysis
Use of
equivalence
scale
Available
breakdowns
Possibility of
measurement
of persistent
poverty
Periodicity of
indicator
published
Type of
data
source
Kazakhstan Expenditure/ consumption expenditure
Monetary and in kind
60% of equivalised median consumption expenditure
Private households
Person National equivalence scale7
Demographic, geographic
No Quarter, annual
Sample survey
Republic of
Moldova
Income/ disposable income
Monetary and in kind
60% of the equivalised median disposable income
Private households
Person Modified OECD scale
Demographic, geographic
Partially Annual Sample survey
Russian
Federation
Income/ disposable income
Monetary and in kind
40%, 50% and 60% of the equivalised median disposable income
Private households
Person Square root scale
Demographic, geographic, social
No Annual Sample survey
Ukraine Expenditure/ consumption expenditure
Monetary and in kind
75% and 60% of equivalised median consumption expenditure
Private households
Household, persons
National equivalence scale8
Demographic, geographic, social
Yes Quarter, annual
Sample survey
Other countries
Bosnia and
Herzegovina
Expenditure/ consumption expenditure
Monetary and in kind
60% of equivalised median consumption expenditure
Private households
Household, person
Modified OECD scale
Demographic, geographic, social
No Every 3‐4 years
Sample survey
Canada Income/
disposable
income
Monetary 50% of the
equivalised
median disposable
income
Private
households
Household Square root
scale
Demographic,
geographic,
social
No Annual Sample
survey
AnnexII
193
Country Poverty
indicator
Welfare
indicator
Poverty Threshold
(line)
Coverage Unit of
analysis
Use of
equivalence
scale
Available
breakdowns
Possibility of
measurement
of persistent
poverty
Periodicity of
indicator
published
Type of
data
source
Israel Income/ disposable income
Monetary 50% of equivalised median disposable income
Private households
Household, persons
National equivalence scale2
Demographic, geographic
No Annual Sample survey
Japan
Income/
disposable
income
Monetary 50% of the
equivalised
median disposable
income
Private
households
Person Square root
scale
Demographic,
social
No Every 3 years Sample
survey
Montenegro Income/
disposable
income
Monetary
and in kind
60% of the
equivalised
median disposable
income
Private
households
Person Modified
OECD scale
Demographic,
geographic,
social
No Annual Sample
survey
New
Zealand
Income/
disposable
income
Monetary 60% of
contemporary
median household
equivalised
Private
households
Person Revised
Jensen Scale
(1988)9
Demographic No Annual Sample
survey
Income/
disposable
income
Monetary 60% of
contemporary
median household
equivalised
disposable income
less 25% as an
allowance for
housing costs
Private
households
Person Revised
Jensen Scale
(1988)
Demographic No Annual Sample
survey
AnnexII
194
Country Poverty
indicator
Welfare
indicator
Poverty Threshold
(line)
Coverage Unit of
analysis
Use of
equivalence
scale
Available
breakdowns
Possibility of
measurement
of persistent
poverty
Periodicity of
indicator
published
Type of
data
source
Turkey Income/ disposable income
Monetary and in kind
40%, 50%, 60% and 70% of the equivalised median disposable income
Private households
Person Modified OECD scale
Geographic Yes Annual Sample survey
Source: National Statistical Offices (2015). Note: 1 First member of a household ‐ 1, each subsequent adult (aged 14+ in household) ‐ 0.66, each child (aged less than 14) ‐ 0.33; 2 First member of a household – 1.25,
second member ‐ 0.75, third member ‐ 0.65, fourth and fifth member ‐ 0.55 each, sixth and seventh member ‐ 0.50 each, eighth member ‐ 0.45, ninth and each subsequent member ‐ 0.40. The official measurement in Israel is performed by the National Insurance Institute (NII); 3 • 1 (N of HH members) ‐ 0.6 (correction factor);• 2 (N of HH members) ‐ 1.0 (correction factor);• 3 (N of HH members) ‐ 1.33 (correction factor);• 4 (N of HH members) ‐ 1.63 (correction factor);• 5 (N of HH members) ‐ 1.90 (correction factor);• 6 (N of HH members) ‐ 2.16 (correction factor);• 7 or more (N of HH members) ‐ 2.40 (correction factor); 4 Do not include student households and households with (annual) incomplete income data; 5 1.0 – the first adult in the household,0.8 – every other adult, 0.9 – every child between 6‐18 years old,0.7 – every child between 3‐5 years, 0.5 – every child under 3 years old; 6 1. Child (aged 0‐7) ‐ 0.64; 2. Adult (aged 8‐15) ‐ 1; 3. Working age male (aged 16‐64) ‐ 1; 4. Working age female (aged 16‐59) ‐ 0.84; 5. Pension age male (aged 65 and more) ‐ 0.88; 6. Pension age female (aged 60 and more) ‐ 0.76. In the power of 0.8 (cohabitation coefficient); 7 First member of a household ‐ 1, each second and subsequent member of a household ‐ 0.8; 8 First member of a household ‐ 1, each second and subsequent member of a household ‐ 0.7; 9 Takes into account number and age of children
AnnexII
195
Table II.3 Subjective poverty Country Type of
indicator
Coverage Unit of analysis Use of
equivalence
scale
Available
breakdowns
Possibility of
measurement of
persistent poverty
Periodicity of
indicator
published
Type of data
source
European Statistical System
Switzerland Subjective assessment of living standards
Private households
Person No Demographic, geographic, social
No Annual Sample survey
Eastern Europe, Caucasus and Central Asia
Armenia Subjective assessment of living standards
Private households
Persons aged 16 years and above
No By objective poverty level
No Annual Sample survey
Belarus Subjective assessment of living standards
Private households
Household No Demographic, geographic
No Quarter, annual Sample survey
Kazakhstan Subjective assessment of living standards
Private households
Person National equivalence scale1
Demographic, geographic
No Annual Sample survey
Ukraine Subjective assessment of living standards
Private households
Household No Demographic, geographic, social
No Annual Sample survey
Other countries
Turkey Subjective assessment of living standards
Private households
Household No Only country level No Annual Sample survey
Source: National Statistical Offices (2015). Note:
1 First member of a household ‐ 1, each second and subsequent member of a household ‐ 0.8.
AnnexII
196
Table II.4 Other approaches towards measuring poverty Country Applied
approach
Short description of applied approach (including criteria adopted,
poverty threshold, etc. as appropriate)
Coverage Unit of
analysis
Available
breakdowns
Periodicity of
indicator
published
Type of
data
source
European Statistical System
Austria Deprivation/ material deprivation
Material deprivation (at least 3 out of 9), severe material deprivation (at least 4 out of 9) according to EU methodology
Private households
Persons Demographic, social
Annual Sample survey
At‐risk‐of‐ poverty or social exclusion
a) At‐risk‐of‐poverty ‐ 60% of the equivalised median disposable income b) Severe material deprivation ‐ severe material deprivation (at least 4 out of 9) according to EU methodology c) Living in a household with very low work intensity ‐ number of persons living in a household having a work intensity below a threshold set at 0.20
Private households
Persons Demographic, social
Annual Sample survey
Bulgaria Deprivation/ material deprivation
Material deprivation (at least 3 out of 9), severe material deprivation (at least 4 out of 9) according to EU methodology
Private households
Households, persons
Demographic, geographic, social
Annual Sample survey
Croatia Deprivation/ material deprivation
Material deprivation (at least 3 out of 9), severe material deprivation (at least 4 out of 9) according to EU methodology
Private households
Person Total population
Annual Sample survey
Cyprus At‐risk‐of‐ poverty or social exclusion
a) At‐risk‐of‐poverty ‐ 60% of the equivalised median disposable income b) Severe material deprivation ‐ severe material deprivation (at least 4 out of 9) according to EU methodology c) Living in a household with very low work intensity ‐ number of persons living in a household having a work intensity below a threshold set at 0.20
Private households
Person Demographic, geographic, social
Annual Sample survey
Czechia Deprivation/ material deprivation
Material deprivation (at least 3 out of 9), severe material deprivation (at least 4 out of 9) according to EU methodology
Private households
Person Demographic, social
Annual Sample survey
AnnexII
197
Country Applied
approach
Short description of applied approach (including criteria adopted,
poverty threshold, etc. as appropriate)
Coverage Unit of
analysis
Available
breakdowns
Periodicity of
indicator
published
Type of
data
source
At‐risk‐of‐poverty or social exclusion
a) At‐risk‐of‐poverty ‐ 60% of the equivalised median disposable income b) Severe material deprivation ‐ severe material deprivation (at least 4 out of 9) according to EU methodology c) Living in a household with very low work intensity ‐ number of persons living in a household having a work intensity below a threshold set at 0.20
Private households
Person Demographic, social
Annual Sample survey
Estonia Deprivation/ material deprivation
Material deprivation (at least 3 out of 9), severe material deprivation (at least 4 out of 9) according to EU methodology
Private households
Person Demographic, geographic, social
Annual Sample survey
Germany At‐risk‐of‐poverty or social exclusion
a) At‐risk‐of‐poverty ‐ 60% of the equivalised median disposable income b) Severe material deprivation ‐ severe material deprivation (at least 4 out of 9) according to EU methodology c) Living in a household with very low work intensity ‐ number of persons living in a household having a work intensity below a threshold set at 0.20
Private households
Person Demographic, social
Annual Sample survey
Deprivation/ material deprivation
Material deprivation (at least 3 out of 9), severe material deprivation (at least 4 out of 9) according to EU methodology
Private households
Person Demographic, social
Annual Sample survey
Hungary Deprivation/ material deprivation
Material deprivation (at least 3 out of 9), severe material deprivation (at least 4 out of 9) according to EU methodology
Private households
Person Demographic Annual Sample survey
AnnexII
198
Country Applied
approach
Short description of applied approach (including criteria adopted,
poverty threshold, etc. as appropriate)
Coverage Unit of
analysis
Available
breakdowns
Periodicity of
indicator
published
Type of
data
source
Ireland Deprivation/ material deprivation
Households that are excluded and marginalised from consuming goods and services which are considered the norm for other people in society, due to an inability to afford them, are considered to be deprived. The identification of the marginalised or deprived is currently achieved on the basis of a set of eleven basic deprivation indicators: 1. Two pairs of strong shoes 2. A warm waterproof overcoat 3. Buy new (not second‐hand) clothes 4. Eat meat with meat, chicken, fish (or vegetarian equivalent) every second day 5. Have a roast joint or its equivalent once a week 6. Had to go without heating during the last year through lack of money 7. Keep the home adequately warm 8. Buy presents for family or friends at least once a year 9. Replace any worn out furniture 10. Have family or friends for a drink or meal once a month 11. Have a morning, afternoon or evening out in the last fortnight for entertainment Individuals who experience two or more of the eleven listed items are considered to be experiencing enforced deprivation. This is the basis for calculating the deprivation rate.
Private households
Person Demographic, geographic, social
Annual Sample survey
Consistent poverty
The consistent poverty measure looks at those persons who are defined as being at risk of poverty and experiencing enforced deprivation (experiencing two or more types of deprivation). An individual is defined as being in ‘consistent poverty’ if they are: • Identified as being at risk of poverty and • Living in a household deprived of two or more of the eleven basic deprivation items listed above.
Private households
Person Demographic, geographic, social
Annual Sample survey
AnnexII
199
Country Applied
approach
Short description of applied approach (including criteria adopted,
poverty threshold, etc. as appropriate)
Coverage Unit of
analysis
Available
breakdowns
Periodicity of
indicator
published
Type of
data
source
Italy Extreme (home‐lessness)
Those who live in: a) public spaces; b) a night‐time dormitory and/or are forced to spend many hours of the day in a public space; c) live in shelters for the homeless/temporary lodgings; lodgings provided in support of those in specific social situations.
Homeless population
Homeless person
Demographic, geographic, social
Occasional (first survey in 2011, follow up in 2014)
Sample survey
At‐risk‐of‐poverty or social exclusion
a) At‐risk‐of‐poverty ‐ 60% of the equivalised median disposable income b) Severe material deprivation ‐ severe material deprivation (at least 4 out of 9) according to EU methodology c) Living in a household with very low work intensity ‐ number of persons living in a household having a work intensity below a threshold set at 0.20
Private households
Household, person
Demographic, geographic, social
Annual Sample survey
Deprivation/ material deprivation
Material deprivation (at least 3 out of 9), severe material deprivation (at least 4 out of 9) according to EU methodology
Private households
Household, person
Demographic, geographic, social
Annual Sample survey
Latvia At‐risk‐of‐poverty or social exclusion
a) At‐risk‐of‐poverty ‐ 60% of the equivalised median disposable income b) Severe material deprivation ‐ severe material deprivation (at least 4 out of 9) according to EU methodology c) Living in a household with very low work intensity ‐ number of persons living in a household having a work intensity below a threshold set at 0.20
Private households
Household, person
Demographic, geographic, social
Annual Sample survey
Deprivation/ material deprivation
Severe material deprivation (at least 4 out of 9) according to EU methodology
Private households
Household, person
Demographic, geographic, social
Annual Sample survey
Low work intensity
Living in a household with very low work intensity ‐ number of persons living in a household having a work intensity below a threshold set at 0.20
Private households
Household, person
Demographic, geographic, social
Annual Sample survey
AnnexII
200
Country Applied
approach
Short description of applied approach (including criteria adopted,
poverty threshold, etc. as appropriate)
Coverage Unit of
analysis
Available
breakdowns
Periodicity of
indicator
published
Type of
data
source
Guaranteed minimum income (GMI) beneficiaries
The GMI benefit amount is calculated as a difference between the guaranteed minimum income level for each family member (established by the Cabinet of Ministers) and the total income of a family (person)
Total population
Family, person
Demographic, geographic, social
Monthly, annual
Administrative records
Needy persons
Needy persons status is granted if the income per capita per month does not reach 128 EUR
Total population
Family, person
Demographic, geographic, social
Monthly, annual
Sample survey
Lithuania Deprivation/ material deprivation
Material deprivation (at least 3 out of 9), severe material deprivation (at least 4 out of 9) according to EU methodology
Private households
Person Demographic, geographic, social
Annual Sample survey
At‐risk‐of‐poverty or social exclusion
a) At‐risk‐of‐poverty ‐ 60% of the equivalised median disposable income b) Severe material deprivation ‐ severe material deprivation (at least 4 out of 9) according to EU methodology c) Living in a household with very low work intensity ‐ number of persons living in a household having a work intensity below a threshold set at 0.20
Private households
Person Demographic Annual Sample survey
Netherlands At‐risk‐of‐poverty or social exclusion
a) At‐risk‐of‐poverty ‐ 60% of the equivalised median disposable income b) Severe material deprivation ‐ severe material deprivation (at least 4 out of 9) according to EU methodology c) Living in a household with very low work intensity ‐ number of persons living in a household having a work intensity below a threshold set at 0.20
Private households
Household, person
Demographic, social
Annual Sample survey
Deprivation/ material deprivation
Material deprivation (at least 3 out of 9), severe material deprivation (at least 4 out of 9) according to EU methodology
Private households
Household, person
Demographic, social
Annual Sample survey
AnnexII
201
Country Applied
approach
Short description of applied approach (including criteria adopted,
poverty threshold, etc. as appropriate)
Coverage Unit of
analysis
Available
breakdowns
Periodicity of
indicator
published
Type of
data
source
Poland Indicator of living conditions poverty
% of households in which at least 10 of 30 symptoms of poor living conditions were observed (symptoms concerning the dwelling quality, the provision of durable consumer goods, and the deprivation of various types of consumer needs (financial and nonfinancial)
Private households
Household Demographic, geographic, social
Every 4‐5 years
Sample survey
Indicator of poverty in terms of the lack of budget balance
% of households in which at least 4 out of 7 symptoms of “inability to deal with their budget” were observed (symptoms concerning both, the subjective opinions of households on their material status, and the facts testifying to budget difficulties faced by the household ‐ including payment arrears)
Private households
Household Demographic, geographic, social
Every 4‐5 years
Sample survey
Multidimen‐sional poverty
% of households affected by the following three forms of poverty at the same time: income poverty, living conditions poverty and poverty in terms of the lack of budget balance
Private households
Household Demographic, geographic, social
Every 4‐5 years
Sample survey
Deprivation/ material deprivation
Material deprivation (at least 3 out of 9), severe material deprivation (at least 4 out of 9) according to EU methodology
Private households
Household Demographic, geographic, social
Annual Sample survey
Slovakia At‐risk‐of‐poverty or social exclusion
a) At‐risk‐of‐poverty ‐ 60% of the equivalised median disposable income b) Severe material deprivation ‐ severe material deprivation (at least 4 out of 9) according to EU methodology c) Living in a household with very low work intensity ‐ number of persons (age group 0‐59) living in a household having a work intensity below a threshold set at 0.20
Private households
Person Demographic, geographic, social
Annual Sample survey
Low work intensity
Living in a household with very low work intensity ‐ number of persons (age group 0‐59) living in a household having a work intensity below a threshold set at 0.20
Private households
Person Demographic Annual Sample survey
Spain Deprivation/ material deprivation
Material deprivation (at least 3 out of 9), severe material deprivation (at least 4 out of 9) according to EU methodology
Private households
Person Demographic, geographic, social
Annual Sample survey
AnnexII
202
Country Applied
approach
Short description of applied approach (including criteria adopted,
poverty threshold, etc. as appropriate)
Coverage Unit of
analysis
Available
breakdowns
Periodicity of
indicator
published
Type of
data
source
At‐risk‐of‐poverty or social exclusion
a) At‐risk‐of‐poverty ‐ 60% of the equivalised median disposable income b) Severe material deprivation ‐ severe material deprivation (at least 4 out of 9) according to EU methodology c) Living in a household with very low work intensity ‐ number of persons living in a household having a work intensity below a threshold set at 0.20
Private households
Person Demographic, geographic, social
Annual Sample survey
Switzerland Deprivation/ material deprivation
Material deprivation (at least 3 out of 9), severe material deprivation (at least 4 out of 9) according to EU methodology
Private households
Person Demographic, geographic, social
Annual Sample survey
United
Kingdom
Deprivation/ material deprivation
Separate measures provided for children and for pensioners based on suites of 21 and 15 questions respectively around access to specific goods and services (informed by independent academic analysis). Responses are prevalence weighted and scaled to give a score out of 100. Children are considered to be in material deprivation if they live in a family with a final score of over 25, for pensioners if the final score is over 20. For children, material deprivation is reported in conjunction with 50% and 70% BHC median income.
Private Households
Person Only country level
Annual Sample survey
Eastern Europe, Caucasus and Central Asia
Belarus Deprivation/ material deprivation
Proportion of people living in households who lack at least 4 material deprivations. National list of material deprivations consists of 12 items.
Private households
Person Demographic, geographic
Аnnual Sample survey
Republic of
Moldova
Deprivation/ material deprivation
Small Area Deprivation Index (SADI) tracks multiple types of deprivation (income, economic, geographic, demographic, health and education) which are then computed into a Multiple Deprivation Index (MDI). All primaria in Republic of Moldova are then assigned a rank between 1 and 843 that reflects their level of deprivation, with 1 representing the most deprived primaria.
843 rural communities/villages
Rural communities/villages
Regions and municipalities
Every second year
Administrative data, census
AnnexII
203
Country Applied
approach
Short description of applied approach (including criteria adopted,
poverty threshold, etc. as appropriate)
Coverage Unit of
analysis
Available
breakdowns
Periodicity of
indicator
published
Type of
data
source
Ukraine Deprivation/ material deprivation
Proportion of households with 4 deprivation out of 18 according to the national methodology The national list of deprivations envisaged the two groups: • economic deprivations: 1) lack of funds to not refuse oneself in the most needed not expensive foodstuffs; 2) lack of funds to update if proved necessary the outwear and footwear for cold season for adults once every 5 years 3) lack of funds to purchase if proved necessary the new clothes and footwear for children; 4) absence of TV‐set; 5) absence of refrigerator; 6) absence of housing under normal conditions (the available housing requires capital repair, it is damp, ramshackle, old); 7) lack of funds for timely and full payments of bills for housing and the necessary services to keep it or pay for gas to cook meals; 8) availability of living floor space that does not exceed 5 sq. m per person; 9) lack of funds to pay for the doctor's needed services (apart from dentists) in medical institution (because of absence or it is difficult to get such services for free), analysis, surveys, procedures prescripted by the doctor; 10) lack of funds to pay for the needed medicines and medical equipment prescripted by the doctor; 11) lack of funds to be treated in the hospital without a surgery operation or vital surgery operation (apart from cosmetic one) and further related treatment in the hospital (because of absence of such services for free); 12) lack of funds to obtain any vocational education; • deprivation by access, i.e. insufficient development of infrastructure as a characteristic of geographical accessibility of services and non‐geographical barriers:
Private households
Household
Demographic, geographic, social
Every two years
Sample survey
AnnexII
204
Country Applied
approach
Short description of applied approach (including criteria adopted,
poverty threshold, etc. as appropriate)
Coverage Unit of
analysis
Available
breakdowns
Periodicity of
indicator
published
Type of
data
source
13) absence of a nearby housing, retail outlets; 14) absence of institutions rendering consumer services in inhabited locality; 15) absence of a nearby housing, medical institution, drug‐store; 16) inability to secure the inhabited locality with timely services of the fast medical aid; 17) absence of a nearby housing, pre‐school institutions; 18) absence of regular daily transport connection with other inhabited locality with more developed infrastructure".
Material deprivation (at least 3 out of 9), severe material deprivation (at least 4 out of 9) according to EU methodology
Private households
Person Demographic, geographic, social
Every two years
Sample survey
Relative poverty + deprivation
Households in which consumption expenditure per equivalent adult is less than 75% of equivalised median consumption expenditure and is deprived by 4 out of 18 attributes
Private households
Household Country wide only
Every two years
Sample survey
Other countries
Bosnia and
Herzegovina
Deprivation/ material deprivation
Material deprivation (at least 3 out of 9), severe material deprivation (at least 4 out of 9) according to EU methodology
Private households
Household, persons
Demographic, geographic, social
Every 3‐4 years
Sample survey
China Multidimen‐sional Poverty
Dimensions: monetary and non‐monetary indicators (more than 30 indicators)
Rural households
Person Demographic Annual Sample survey
AnnexII
205
Country Applied
approach
Short description of applied approach (including criteria adopted,
poverty threshold, etc. as appropriate)
Coverage Unit of
analysis
Available
breakdowns
Periodicity of
indicator
published
Type of
data
source
Colombia Multidimen‐sional Poverty
Dimensions: 1. Education of household members 2. Childhood and youth conditions 3. Health 4. Employment 5. Access to household utilities and living conditions. A household is considered as poor if it has deprivation in at least 33% of the indicators, taking into account the weight of each indicator. This index has a nested weighting structure where each dimension is equally weighted (0.2), and inside the dimension, each indicator has the same weight.
Private households
Household Demographic Annual Sample survey
Mexico Multidimen‐sional poverty
Dimensions: income, social deprivations Private households
Person Demographic, geographic, social
Every 2 years for federal entities; every 5 years for municipalities
Sample survey
AnnexII
206
Country Applied
approach
Short description of applied approach (including criteria adopted,
poverty threshold, etc. as appropriate)
Coverage Unit of
analysis
Available
breakdowns
Periodicity of
indicator
published
Type of
data
source
New
Zealand
Hardship A 40‐item Economic Living Standards Index (ELSI) ranks households from low to high living standards using non‐monetary indicators (NMIs). To create the ELSI scores, the NMI items are scored from two different perspectives: 1. from an enforced lack perspective in which respondents do not have essential items because of the cost, or have to severely cut back on purchases because the money is needed for other essentials: for example, unable (because of the cost) to have regular good meals, two pairs of shoes in good repair for everyday activities, or visit the doctor; putting up with the cold, and so on because money is needed for other basics) 2. from the perspective of the degree of restriction/freedom reported for having or purchasing desirable non‐essentials (while having the essentials) – a freedoms enjoyed perspective, for short: for example, having all the essentials, and in addition not having to cut back on local trips, not having to put off replacing broken or worn out appliances, being able to take an overseas holiday every three years or so if desired, and not having any great restrictions on purchasing clothing. A state of hardship (unacceptably low material well‐being) is characterised by having many enforced lacks of essentials and few or no freedoms. Higher living standards are characterised by having all the essentials (no enforced lacks) and also having many freedoms and few restrictions in relation to the non‐essential items that are asked about. The ELSI hardship threshold is set to be equivalent to one set at 6 or more deprivations out of 16 in the calibration list.
Private households
Person Demographic Annual Sample survey
AnnexII
207
Country Applied
approach
Short description of applied approach (including criteria adopted,
poverty threshold, etc. as appropriate)
Coverage Unit of
analysis
Available
breakdowns
Periodicity of
indicator
published
Type of
data
source
Hardship The Material Wellbeing Index is a revised and updated version of ELSI, building off what has been learnt from using ELSI over the last decade. There are a set of 24 items that go into constructing the index and these items cover 6 areas: 1. Ownership (have, don’t have and enforced lack) 2. Social participation (do, don’t do and enforced lack) 3. Economising (not at all, a little, a lot) – to keep down costs to help in paying for (other) basic items 4. Freedoms/Restrictions (buying items) 5. Financial strain (in last 12 months) 6. Housing problems (no problem, minor problem, major problem)
Private households
Person Demographic Annual Sample survey
Source: National Statistical Offices (2015).